3x-ui/install.sh

2055 lines
93 KiB
Bash
Raw Normal View History

2023-02-09 11:18:06 -08:00
#!/bin/bash
red='\033[0;31m'
green='\033[0;32m'
blue='\033[0;34m'
2023-02-09 11:18:06 -08:00
yellow='\033[0;33m'
plain='\033[0m'
xui_folder="${XUI_MAIN_FOLDER:=/usr/local/x-ui}"
xui_service="${XUI_SERVICE:=/etc/systemd/system}"
2023-02-09 11:18:06 -08:00
# check root
2026-05-19 04:09:35 -07:00
[[ $EUID -ne 0 ]] && echo -e "${red}Fatal error: ${plain} Please run this script with root privilege \n " && exit 1
2023-02-09 11:18:06 -08:00
2023-03-11 07:05:35 -08:00
# Check OS and set release variable
if [[ -f /etc/os-release ]]; then
source /etc/os-release
release=$ID
2026-05-04 04:20:24 -07:00
elif [[ -f /usr/lib/os-release ]]; then
2023-03-11 07:05:35 -08:00
source /usr/lib/os-release
release=$ID
2023-02-09 11:18:06 -08:00
else
2026-05-19 04:09:35 -07:00
echo "Failed to check the system OS, please contact the author!" >&2
exit 1
2023-02-09 11:18:06 -08:00
fi
2023-03-11 07:05:35 -08:00
echo "The OS release is: $release"
2023-02-09 11:18:06 -08:00
2024-04-01 01:42:28 -07:00
arch() {
case "$(uname -m)" in
x86_64 | x64 | amd64) echo 'amd64' ;;
i*86 | x86) echo '386' ;;
armv8* | armv8 | arm64 | aarch64) echo 'arm64' ;;
armv7* | armv7 | arm) echo 'armv7' ;;
armv6* | armv6) echo 'armv6' ;;
armv5* | armv5) echo 'armv5' ;;
s390x) echo 's390x' ;;
*) echo -e "${green}Unsupported CPU architecture! ${plain}" && rm -f "$(realpath "$0")" && exit 1 ;;
esac
}
2024-01-20 01:32:35 -08:00
2025-03-21 08:30:31 -07:00
echo "Arch: $(arch)"
2023-02-09 11:18:06 -08:00
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
# Non-interactive mode: triggered explicitly via XUI_NONINTERACTIVE=1, or
# implicitly when stdin is not a TTY (e.g. `curl ... | bash`, cloud-init).
# In this mode every prompt below is replaced by an env var or a sane default.
if [[ "${XUI_NONINTERACTIVE:-0}" == "1" ]] || [[ ! -t 0 ]]; then
NONINTERACTIVE=1
else
NONINTERACTIVE=0
fi
export NONINTERACTIVE
2025-12-27 15:03:33 -08:00
# Simple helpers
is_ipv4() {
[[ "$1" =~ ^([0-9]{1,3}\.){3}[0-9]{1,3}$ ]] && return 0 || return 1
}
is_ipv6() {
[[ "$1" =~ : ]] && return 0 || return 1
}
is_ip() {
is_ipv4 "$1" || is_ipv6 "$1"
}
is_domain() {
[[ "$1" =~ ^([A-Za-z0-9](-*[A-Za-z0-9])*\.)+(xn--[a-z0-9]{2,}|[A-Za-z]{2,})$ ]] && return 0 || return 1
}
# acme.sh's standalone server binds IPv4 by default; --listen-v6 makes it
# v6-only, which breaks HTTP-01 validation when the domain's A record points
# at this host's IPv4 (#4994). Only force IPv6 when the host has no global
# IPv4 address at all.
acme_listen_flag() {
if ip -4 addr show scope global 2> /dev/null | grep -q "inet "; then
echo ""
else
echo "--listen-v6"
fi
}
# Port helpers
is_port_in_use() {
local port="$1"
2026-05-04 04:20:24 -07:00
if command -v ss > /dev/null 2>&1; then
ss -ltn 2> /dev/null | awk -v p=":${port}$" '$4 ~ p {exit 0} END {exit 1}'
return
fi
2026-05-04 04:20:24 -07:00
if command -v netstat > /dev/null 2>&1; then
netstat -lnt 2> /dev/null | awk -v p=":${port} " '$4 ~ p {exit 0} END {exit 1}'
return
fi
2026-05-04 04:20:24 -07:00
if command -v lsof > /dev/null 2>&1; then
lsof -nP -iTCP:${port} -sTCP:LISTEN > /dev/null 2>&1 && return 0
fi
return 1
2025-12-27 15:03:33 -08:00
}
2026-05-19 04:09:35 -07:00
install_base() {
case "${release}" in
ubuntu | debian | armbian)
apt-get update && apt-get install -y -q cron curl tar tzdata socat ca-certificates openssl
;;
fedora | amzn | virtuozzo | rhel | almalinux | rocky | ol)
dnf makecache -y && dnf install -y -q cronie curl tar tzdata socat ca-certificates openssl
2026-05-19 04:09:35 -07:00
;;
centos)
if [[ "${VERSION_ID}" =~ ^7 ]]; then
yum makecache -y && yum install -y cronie curl tar tzdata socat ca-certificates openssl
2026-05-19 04:09:35 -07:00
else
dnf makecache -y && dnf install -y -q cronie curl tar tzdata socat ca-certificates openssl
2026-05-19 04:09:35 -07:00
fi
;;
arch | manjaro | parch)
pacman -Sy --noconfirm cronie curl tar tzdata socat ca-certificates openssl
2026-05-19 04:09:35 -07:00
;;
opensuse-tumbleweed | opensuse-leap)
zypper refresh && zypper -q install -y cron curl tar timezone socat ca-certificates openssl
;;
alpine)
apk update && apk add dcron curl tar tzdata socat ca-certificates openssl
;;
*)
apt-get update && apt-get install -y -q cron curl tar tzdata socat ca-certificates openssl
;;
esac
}
2024-06-24 05:43:39 -07:00
gen_random_string() {
local length="$1"
2026-05-04 04:20:24 -07:00
openssl rand -base64 $((length * 2)) \
| tr -dc 'a-zA-Z0-9' \
| head -c "$length"
2024-06-24 05:43:39 -07:00
}
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
# prompt_or_default VARNAME "prompt text" "default" [ENV_NAME]
# Interactive: read into VARNAME. Non-interactive: VARNAME = ${ENV_NAME:-default}.
# ENV_NAME defaults to VARNAME when omitted. Keeps every interactive prompt
# string byte-for-byte identical to the original `read -rp`.
prompt_or_default() {
local __var="$1" __prompt="$2" __default="$3" __env="${4:-$1}"
if [[ "$NONINTERACTIVE" == "1" ]]; then
printf -v "$__var" '%s' "${!__env:-$__default}"
else
# shellcheck disable=SC2229
read -rp "$__prompt" "$__var"
fi
}
# write_install_result <user> <pass> <port> <webpath> <scheme> <host> <token> <dbtype>
# Persists a parseable, root-only credentials file consumed by cloud-init/MOTD.
# Values are written with printf '%q' so a pinned password/username containing
# spaces, quotes, $(...) or backticks is shell-escaped and the file stays safely
# source-able (consumers do '. install-result.env'). For the alphanumeric random
# values gen_random_string emits, %q is a no-op. This is a DIFFERENT file from the
# Postgres env file (/etc/default/x-ui).
write_install_result() {
local u="$1" p="$2" port="$3" wbp="$4" scheme="$5" host="$6" token="$7" dbtype="$8"
local result_file="/etc/x-ui/install-result.env"
local url_host="${host:-SERVER_IP_UNKNOWN}"
install -d -m 700 /etc/x-ui 2> /dev/null
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
local prev_umask
prev_umask=$(umask)
umask 077
if ! {
printf 'XUI_USERNAME=%q\n' "$u"
printf 'XUI_PASSWORD=%q\n' "$p"
printf 'XUI_PANEL_PORT=%q\n' "$port"
printf 'XUI_WEB_BASE_PATH=%q\n' "$wbp"
printf 'XUI_ACCESS_URL=%q\n' "${scheme}://${url_host}:${port}/${wbp}"
printf 'XUI_API_TOKEN=%q\n' "$token"
printf 'XUI_DB_TYPE=%q\n' "$dbtype"
} > "$result_file"; then
umask "$prev_umask"
echo -e "${yellow}Warning: failed to write ${result_file}.${plain}" >&2
return 1
fi
umask "$prev_umask"
chmod 600 "$result_file" 2> /dev/null
chown root:root "$result_file" 2> /dev/null || true
echo -e "${green}Install result written to ${result_file} (mode 600).${plain}"
}
# RHEL-family initdb writes pg_hba.conf host rules with ident auth, which
# compares the OS username against the Postgres role and always rejects the
# randomly generated panel role over TCP (#5806). Prepend password-auth rules
# scoped to the panel database; first match wins, and md5 also accepts
# scram-sha-256-stored verifiers, so this works on every supported distro.
pg_ensure_hba_password_auth() {
local pg_db="$1"
local hba_file
hba_file=$(sudo -u postgres psql -tAc 'SHOW hba_file' 2> /dev/null | tr -d '[:space:]')
[[ -n "${hba_file}" && -f "${hba_file}" ]] || return 0
grep -Eq "^host[[:space:]]+${pg_db}[[:space:]]" "${hba_file}" && return 0
local tmp
tmp=$(mktemp) || return 1
{
echo "# Added by 3x-ui: allow password logins for the panel database."
echo "host ${pg_db} all 127.0.0.1/32 md5"
echo "host ${pg_db} all ::1/128 md5"
cat "${hba_file}"
} > "${tmp}" || {
rm -f "${tmp}"
return 1
}
cat "${tmp}" > "${hba_file}" || {
rm -f "${tmp}"
return 1
}
rm -f "${tmp}"
sudo -u postgres psql -tAc 'SELECT pg_reload_conf()' > /dev/null 2>&1 || true
}
# New installs get this PostgreSQL major from the official PGDG repository; distros or
# CPU architectures PGDG does not build for fall back to the distribution's own package.
PG_TARGET_MAJOR=18
pg_has_debian_cluster() {
command -v pg_lsclusters > /dev/null 2>&1 && [[ -n "$(pg_lsclusters -h 2> /dev/null)" ]]
}
# Installs postgresql-${PG_TARGET_MAJOR} from apt.postgresql.org. On failure it removes the
# repository it added, so an unsupported release cannot break later `apt-get update` runs.
pg_apt_install_pgdg_server() {
local list added_lists=()
for list in /etc/apt/sources.list.d/pgdg.list /etc/apt/sources.list.d/pgdg.sources; do
[[ -e "${list}" ]] || added_lists+=("${list}")
done
if apt-get install -y -q postgresql-common ca-certificates gnupg >&2 \
&& /usr/share/postgresql-common/pgdg/apt.postgresql.org.sh -y >&2 \
&& apt-get install -y -q "postgresql-${PG_TARGET_MAJOR}" >&2; then
return 0
fi
for list in "${added_lists[@]}"; do
rm -f "${list}"
done
apt-get update >&2 || true
echo -e "${yellow}PostgreSQL ${PG_TARGET_MAJOR} is not available for this system; installing the distribution's PostgreSQL instead.${plain}" >&2
return 1
}
# Installs postgresql${PG_TARGET_MAJOR}-server from the PGDG yum repository (EL 8+ only); on
# failure it undoes the repository and module changes it made. The caller runs initdb.
pg_el_install_pgdg_server() {
local elver
elver=$(rpm -E %rhel 2> /dev/null)
[[ "${elver}" =~ ^[0-9]+$ && "${elver}" -ge 8 ]] && command -v dnf > /dev/null 2>&1 || return 1
[[ -f /var/lib/pgsql/data/PG_VERSION ]] && return 1
local added_repo=false
if ! rpm -q pgdg-redhat-repo > /dev/null 2>&1; then
if dnf install -y -q "https://download.postgresql.org/pub/repos/yum/reporpms/EL-${elver}-$(uname -m)/pgdg-redhat-repo-latest.noarch.rpm" >&2; then
added_repo=true
dnf -qy module disable postgresql > /dev/null 2>&1 || true
fi
fi
if rpm -q pgdg-redhat-repo > /dev/null 2>&1 \
&& dnf install -y -q "postgresql${PG_TARGET_MAJOR}-server" "postgresql${PG_TARGET_MAJOR}-contrib" >&2; then
return 0
fi
if [[ "${added_repo}" == "true" ]]; then
dnf -qy module reset postgresql > /dev/null 2>&1 || true
dnf remove -y -q pgdg-redhat-repo >&2 || true
fi
echo -e "${yellow}PostgreSQL ${PG_TARGET_MAJOR} is not available for this system; installing the distribution's PostgreSQL instead.${plain}" >&2
return 1
}
install_postgres_local() {
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
local pg_user pg_pass
pg_pass=$(gen_random_string 24)
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
local pg_db="xui"
local pg_host="127.0.0.1"
local pg_port="5432"
local pg_service="postgresql"
case "${release}" in
ubuntu | debian | armbian)
apt-get update >&2 || return 1
if pg_has_debian_cluster || ! pg_apt_install_pgdg_server; then
apt-get install -y -q postgresql >&2 || return 1
fi
;;
fedora | amzn | virtuozzo | rhel | almalinux | rocky | ol | centos)
if [[ "${release}" != "fedora" && "${release}" != "amzn" ]] && pg_el_install_pgdg_server; then
pg_service="postgresql-${PG_TARGET_MAJOR}"
[[ -f "/var/lib/pgsql/${PG_TARGET_MAJOR}/data/PG_VERSION" ]] \
|| "/usr/pgsql-${PG_TARGET_MAJOR}/bin/postgresql-${PG_TARGET_MAJOR}-setup" initdb >&2 || return 1
else
if [[ "${release}" == "centos" && "${VERSION_ID}" =~ ^7 ]]; then
yum install -y postgresql-server postgresql-contrib >&2 || return 1
else
dnf install -y -q postgresql-server postgresql-contrib >&2 || return 1
fi
[[ -d /var/lib/pgsql/data && -f /var/lib/pgsql/data/PG_VERSION ]] || postgresql-setup --initdb >&2 || return 1
fi
;;
arch | manjaro | parch)
pacman -Sy --noconfirm postgresql >&2 || return 1
if [[ ! -f /var/lib/postgres/data/PG_VERSION ]]; then
sudo -u postgres initdb -D /var/lib/postgres/data >&2 || return 1
fi
;;
opensuse-tumbleweed | opensuse-leap)
zypper -q install -y postgresql-server postgresql-contrib >&2 || return 1
if [[ ! -f /var/lib/pgsql/data/PG_VERSION ]]; then
install -d -o postgres -g postgres -m 700 /var/lib/pgsql/data >&2 || return 1
su - postgres -c "initdb -D /var/lib/pgsql/data" >&2 || return 1
fi
;;
alpine)
apk add --no-cache postgresql postgresql-contrib >&2 || return 1
if [[ ! -f /var/lib/postgresql/data/PG_VERSION ]]; then
/etc/init.d/postgresql setup >&2 || return 1
fi
rc-update add postgresql default >&2 2> /dev/null || true
rc-service postgresql start >&2 || return 1
;;
*)
echo -e "${red}Unsupported distro for automatic PostgreSQL install: ${release}${plain}" >&2
return 1
;;
esac
if [[ "${release}" != "alpine" ]]; then
systemctl enable --now "${pg_service}" >&2 || return 1
fi
# Wait briefly for the server to accept connections.
local i
for i in 1 2 3 4 5; do
sudo -u postgres psql -tAc 'SELECT 1' > /dev/null 2>&1 && break
sleep 1
done
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
local existing_owner=""
existing_owner=$(sudo -u postgres psql -tAc \
"SELECT pg_catalog.pg_get_userbyid(datdba) FROM pg_database WHERE datname='${pg_db}'" 2> /dev/null \
| tr -d '[:space:]')
if [[ -n "${existing_owner}" && "${existing_owner}" != "postgres" ]]; then
pg_user="${existing_owner}"
else
pg_user=$(gen_random_string 8)
fi
# Idempotent role/db creation. Identifiers are double-quoted because a
# random username may start with a digit, which Postgres rejects unquoted.
sudo -u postgres psql -tAc "SELECT 1 FROM pg_roles WHERE rolname='${pg_user}'" 2> /dev/null \
| grep -q 1 \
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
|| sudo -u postgres psql -c "CREATE USER \"${pg_user}\" WITH PASSWORD '${pg_pass}';" >&2 || return 1
sudo -u postgres psql -tAc "SELECT 1 FROM pg_database WHERE datname='${pg_db}'" 2> /dev/null \
| grep -q 1 \
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
|| sudo -u postgres psql -c "CREATE DATABASE \"${pg_db}\" OWNER \"${pg_user}\";" >&2 || return 1
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
sudo -u postgres psql -c "ALTER USER \"${pg_user}\" WITH PASSWORD '${pg_pass}';" >&2 || return 1
pg_ensure_hba_password_auth "${pg_db}" \
|| echo -e "${yellow}Warning: could not update pg_hba.conf; PostgreSQL may reject the panel's TCP login (ident auth).${plain}" >&2
local pg_pass_enc
pg_pass_enc=$(printf '%s' "${pg_pass}" | sed -e 's/%/%25/g' -e 's/:/%3A/g' -e 's/@/%40/g' -e 's|/|%2F|g' -e 's/?/%3F/g' -e 's/#/%23/g')
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
if [[ -n "${PG_CRED_FILE:-}" ]]; then
local prev_umask
prev_umask=$(umask)
umask 077
if ! cat > "${PG_CRED_FILE}" << EOF; then
PG_USER=${pg_user}
PG_PASS=${pg_pass}
PG_HOST=${pg_host}
PG_PORT=${pg_port}
PG_DB=${pg_db}
EOF
umask "${prev_umask}"
echo -e "${red}Failed to write PostgreSQL credentials to ${PG_CRED_FILE}${plain}" >&2
return 1
fi
umask "${prev_umask}"
fi
echo "postgres://${pg_user}:${pg_pass_enc}@${pg_host}:${pg_port}/${pg_db}?sslmode=disable"
return 0
}
ensure_pg_client() {
if command -v pg_dump > /dev/null 2>&1 && command -v pg_restore > /dev/null 2>&1; then
return 0
fi
echo -e "${yellow}Installing PostgreSQL client tools (pg_dump/pg_restore) for in-panel backup...${plain}" >&2
case "${release}" in
ubuntu | debian | armbian)
apt-get update >&2 && apt-get install -y -q postgresql-client >&2 || return 1
;;
fedora | amzn | virtuozzo | rhel | almalinux | rocky | ol)
dnf install -y -q postgresql >&2 || return 1
;;
centos)
if [[ "${VERSION_ID}" =~ ^7 ]]; then
yum install -y postgresql >&2 || return 1
else
dnf install -y -q postgresql >&2 || return 1
fi
;;
arch | manjaro | parch)
pacman -Sy --noconfirm postgresql >&2 || return 1
;;
opensuse-tumbleweed | opensuse-leap)
zypper -q install -y postgresql >&2 || return 1
;;
alpine)
apk add --no-cache postgresql-client >&2 || return 1
;;
*)
return 1
;;
esac
command -v pg_dump > /dev/null 2>&1 && command -v pg_restore > /dev/null 2>&1
}
2025-12-27 15:03:33 -08:00
install_acme() {
echo -e "${green}Installing acme.sh for SSL certificate management...${plain}"
cd ~ || return 1
2026-05-04 04:20:24 -07:00
curl -s https://get.acme.sh | sh > /dev/null 2>&1
2025-12-27 15:03:33 -08:00
if [ $? -ne 0 ]; then
echo -e "${red}Failed to install acme.sh${plain}"
return 1
else
echo -e "${green}acme.sh installed successfully${plain}"
fi
return 0
}
setup_ssl_certificate() {
local domain="$1"
local server_ip="$2"
local existing_port="$3"
local existing_webBasePath="$4"
2026-05-04 04:20:24 -07:00
2025-12-27 15:03:33 -08:00
echo -e "${green}Setting up SSL certificate...${plain}"
2026-05-04 04:20:24 -07:00
2025-12-27 15:03:33 -08:00
# Check if acme.sh is installed
2026-05-04 04:20:24 -07:00
if ! command -v ~/.acme.sh/acme.sh &> /dev/null; then
2025-12-27 15:03:33 -08:00
install_acme
if [ $? -ne 0 ]; then
echo -e "${yellow}Failed to install acme.sh, skipping SSL setup${plain}"
return 1
fi
fi
2026-05-04 04:20:24 -07:00
2025-12-27 15:03:33 -08:00
# Create certificate directory
local certPath="/root/cert/${domain}"
mkdir -p "$certPath"
2026-05-04 04:20:24 -07:00
2025-12-27 15:03:33 -08:00
# Issue certificate
echo -e "${green}Issuing SSL certificate for ${domain}...${plain}"
echo -e "${yellow}Note: Port 80 must be open and accessible from the internet${plain}"
2026-05-04 04:20:24 -07:00
~/.acme.sh/acme.sh --set-default-ca --server letsencrypt --force > /dev/null 2>&1
~/.acme.sh/acme.sh --issue -d ${domain} $(acme_listen_flag) --standalone --httpport 80 --force
2026-05-04 04:20:24 -07:00
2025-12-27 15:03:33 -08:00
if [ $? -ne 0 ]; then
echo -e "${yellow}Failed to issue certificate for ${domain}${plain}"
echo -e "${yellow}Please ensure port 80 is open and try again later with: x-ui${plain}"
rm -rf ~/.acme.sh/${domain} ~/.acme.sh/${domain}_ecc 2> /dev/null
2026-05-04 04:20:24 -07:00
rm -rf "$certPath" 2> /dev/null
2025-12-27 15:03:33 -08:00
return 1
fi
2026-05-04 04:20:24 -07:00
2025-12-27 15:03:33 -08:00
# Install certificate
~/.acme.sh/acme.sh --installcert --force -d ${domain} \
2025-12-27 15:03:33 -08:00
--key-file /root/cert/${domain}/privkey.pem \
--fullchain-file /root/cert/${domain}/fullchain.pem \
2026-05-04 04:20:24 -07:00
--reloadcmd "systemctl restart x-ui" > /dev/null 2>&1
2025-12-27 15:03:33 -08:00
if [ $? -ne 0 ]; then
echo -e "${yellow}Failed to install certificate${plain}"
return 1
fi
2026-05-04 04:20:24 -07:00
2025-12-27 15:03:33 -08:00
# Enable auto-renew
2026-05-04 04:20:24 -07:00
~/.acme.sh/acme.sh --upgrade --auto-upgrade > /dev/null 2>&1
2026-05-19 04:09:35 -07:00
# Secure permissions: private key readable only by owner
2026-05-04 04:20:24 -07:00
chmod 600 $certPath/privkey.pem 2> /dev/null
chmod 644 $certPath/fullchain.pem 2> /dev/null
2025-12-27 15:03:33 -08:00
# Set certificate for panel
local webCertFile="/root/cert/${domain}/fullchain.pem"
local webKeyFile="/root/cert/${domain}/privkey.pem"
2026-05-04 04:20:24 -07:00
2025-12-27 15:03:33 -08:00
if [[ -f "$webCertFile" && -f "$webKeyFile" ]]; then
2026-05-04 04:20:24 -07:00
${xui_folder}/x-ui cert -webCert "$webCertFile" -webCertKey "$webKeyFile" > /dev/null 2>&1
2025-12-27 15:03:33 -08:00
echo -e "${green}SSL certificate installed and configured successfully!${plain}"
return 0
else
echo -e "${yellow}Certificate files not found${plain}"
return 1
fi
}
2026-01-04 20:28:02 -08:00
# Issue Let's Encrypt IP certificate with shortlived profile (~6 days validity)
# Requires acme.sh and port 80 open for HTTP-01 challenge
setup_ip_certificate() {
local ipv4="$1"
2026-05-04 04:20:24 -07:00
local ipv6="$2" # optional
2025-12-27 15:03:33 -08:00
2026-01-04 20:28:02 -08:00
echo -e "${green}Setting up Let's Encrypt IP certificate (shortlived profile)...${plain}"
echo -e "${yellow}Note: IP certificates are valid for ~6 days and will auto-renew.${plain}"
echo -e "${yellow}Default listener is port 80. If you choose another port, ensure external port 80 forwards to it.${plain}"
2025-12-27 15:03:33 -08:00
2026-01-04 20:28:02 -08:00
# Check for acme.sh
2026-05-04 04:20:24 -07:00
if ! command -v ~/.acme.sh/acme.sh &> /dev/null; then
2026-01-04 20:28:02 -08:00
install_acme
if [ $? -ne 0 ]; then
echo -e "${red}Failed to install acme.sh${plain}"
return 1
fi
2025-12-27 15:03:33 -08:00
fi
2026-01-04 20:28:02 -08:00
# Validate IP address
if [[ -z "$ipv4" ]]; then
echo -e "${red}IPv4 address is required${plain}"
2025-12-27 15:03:33 -08:00
return 1
fi
2026-01-04 20:28:02 -08:00
if ! is_ipv4 "$ipv4"; then
echo -e "${red}Invalid IPv4 address: $ipv4${plain}"
return 1
fi
# Create certificate directory
local certDir="/root/cert/ip"
mkdir -p "$certDir"
# Build domain arguments
local domain_args="-d ${ipv4}"
if [[ -n "$ipv6" ]] && is_ipv6 "$ipv6"; then
domain_args="${domain_args} -d ${ipv6}"
echo -e "${green}Including IPv6 address: ${ipv6}${plain}"
fi
2026-05-19 04:09:35 -07:00
# Set reload command for auto-renewal (add || true so it doesn't fail during first install)
2026-01-04 20:28:02 -08:00
local reloadCmd="systemctl restart x-ui 2>/dev/null || rc-service x-ui restart 2>/dev/null || true"
# Choose port for HTTP-01 listener (default 80, prompt override)
local WebPort=""
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
prompt_or_default WebPort "Port to use for ACME HTTP-01 listener (default 80): " "80" XUI_ACME_HTTP_PORT
WebPort="${WebPort:-80}"
if ! [[ "${WebPort}" =~ ^[0-9]+$ ]] || ((WebPort < 1 || WebPort > 65535)); then
echo -e "${red}Invalid port provided. Falling back to 80.${plain}"
WebPort=80
fi
echo -e "${green}Using port ${WebPort} for standalone validation.${plain}"
if [[ "${WebPort}" -ne 80 ]]; then
echo -e "${yellow}Reminder: Let's Encrypt still connects on port 80; forward external port 80 to ${WebPort}.${plain}"
fi
# Ensure chosen port is available
while true; do
if is_port_in_use "${WebPort}"; then
2026-05-19 04:09:35 -07:00
echo -e "${yellow}Port ${WebPort} is in use.${plain}"
local alt_port=""
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ "$NONINTERACTIVE" == "1" ]]; then
echo -e "${red}Port ${WebPort} is busy; cannot proceed in non-interactive mode.${plain}"
return 1
fi
read -rp "Enter another port for acme.sh standalone listener (leave empty to abort): " alt_port
alt_port="${alt_port// /}"
if [[ -z "${alt_port}" ]]; then
echo -e "${red}Port ${WebPort} is busy; cannot proceed.${plain}"
return 1
fi
if ! [[ "${alt_port}" =~ ^[0-9]+$ ]] || ((alt_port < 1 || alt_port > 65535)); then
echo -e "${red}Invalid port provided.${plain}"
return 1
fi
WebPort="${alt_port}"
continue
else
echo -e "${green}Port ${WebPort} is free and ready for standalone validation.${plain}"
break
fi
done
2026-01-04 20:28:02 -08:00
# Issue certificate with shortlived profile
echo -e "${green}Issuing IP certificate for ${ipv4}...${plain}"
2026-05-04 04:20:24 -07:00
~/.acme.sh/acme.sh --set-default-ca --server letsencrypt --force > /dev/null 2>&1
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
[[ -n "${XUI_ACME_EMAIL:-}" ]] && ~/.acme.sh/acme.sh --register-account -m "${XUI_ACME_EMAIL}" > /dev/null 2>&1
2026-05-04 04:20:24 -07:00
2026-01-04 20:28:02 -08:00
~/.acme.sh/acme.sh --issue \
${domain_args} \
--standalone \
--server letsencrypt \
--certificate-profile shortlived \
--days 6 \
--httpport ${WebPort} \
2026-01-04 20:28:02 -08:00
--force
if [ $? -ne 0 ]; then
echo -e "${red}Failed to issue IP certificate${plain}"
echo -e "${yellow}Please ensure port ${WebPort} is reachable (or forwarded from external port 80)${plain}"
2026-01-04 20:28:02 -08:00
# Cleanup acme.sh data for both IPv4 and IPv6 if specified
rm -rf ~/.acme.sh/${ipv4} ~/.acme.sh/${ipv4}_ecc 2> /dev/null
[[ -n "$ipv6" ]] && rm -rf ~/.acme.sh/${ipv6} ~/.acme.sh/${ipv6}_ecc 2> /dev/null
2026-05-04 04:20:24 -07:00
rm -rf ${certDir} 2> /dev/null
2026-01-04 20:28:02 -08:00
return 1
fi
echo -e "${green}Certificate issued successfully, installing...${plain}"
# Install certificate
# Note: acme.sh may report "Reload error" and exit non-zero if reloadcmd fails,
# but the cert files are still installed. We check for files instead of exit code.
~/.acme.sh/acme.sh --installcert --force -d ${ipv4} \
2026-01-04 20:28:02 -08:00
--key-file "${certDir}/privkey.pem" \
--fullchain-file "${certDir}/fullchain.pem" \
--reloadcmd "${reloadCmd}" 2>&1 || true
# Verify certificate files exist (don't rely on exit code - reloadcmd failure causes non-zero)
if [[ ! -f "${certDir}/fullchain.pem" || ! -f "${certDir}/privkey.pem" ]]; then
echo -e "${red}Certificate files not found after installation${plain}"
# Cleanup acme.sh data for both IPv4 and IPv6 if specified
rm -rf ~/.acme.sh/${ipv4} ~/.acme.sh/${ipv4}_ecc 2> /dev/null
[[ -n "$ipv6" ]] && rm -rf ~/.acme.sh/${ipv6} ~/.acme.sh/${ipv6}_ecc 2> /dev/null
2026-05-04 04:20:24 -07:00
rm -rf ${certDir} 2> /dev/null
2026-01-04 20:28:02 -08:00
return 1
fi
2026-05-04 04:20:24 -07:00
2026-01-04 20:28:02 -08:00
echo -e "${green}Certificate files installed successfully${plain}"
# Enable auto-upgrade for acme.sh (ensures cron job runs)
2026-05-04 04:20:24 -07:00
~/.acme.sh/acme.sh --upgrade --auto-upgrade > /dev/null 2>&1
2026-01-04 20:28:02 -08:00
2026-05-19 04:09:35 -07:00
# Secure permissions: private key readable only by owner
2026-05-04 04:20:24 -07:00
chmod 600 ${certDir}/privkey.pem 2> /dev/null
chmod 644 ${certDir}/fullchain.pem 2> /dev/null
2026-01-04 20:28:02 -08:00
# Configure panel to use the certificate
echo -e "${green}Setting certificate paths for the panel...${plain}"
${xui_folder}/x-ui cert -webCert "${certDir}/fullchain.pem" -webCertKey "${certDir}/privkey.pem"
2026-05-19 04:09:35 -07:00
2026-01-04 20:28:02 -08:00
if [ $? -ne 0 ]; then
2026-05-19 04:09:35 -07:00
echo -e "${yellow}Warning: Could not set certificate paths automatically${plain}"
echo -e "${yellow}Certificate files are at:${plain}"
echo -e " Cert: ${certDir}/fullchain.pem"
echo -e " Key: ${certDir}/privkey.pem"
2026-01-04 20:28:02 -08:00
else
2026-05-19 04:09:35 -07:00
echo -e "${green}Certificate paths configured successfully${plain}"
2026-01-04 20:28:02 -08:00
fi
echo -e "${green}IP certificate installed and configured successfully!${plain}"
echo -e "${green}Certificate valid for ~6 days, auto-renews via acme.sh cron job.${plain}"
2026-05-19 04:09:35 -07:00
echo -e "${yellow}acme.sh will automatically renew and reload x-ui before expiry.${plain}"
2025-12-27 15:03:33 -08:00
return 0
}
# Comprehensive manual SSL certificate issuance via acme.sh
ssl_cert_issue() {
local existing_webBasePath=$(${xui_folder}/x-ui setting -show true | grep 'webBasePath:' | awk -F': ' '{print $2}' | tr -d '[:space:]' | sed 's#^/##')
local existing_port=$(${xui_folder}/x-ui setting -show true | grep 'port:' | awk -F': ' '{print $2}' | tr -d '[:space:]')
2026-05-04 04:20:24 -07:00
2025-12-27 15:03:33 -08:00
# check for acme.sh first
2026-05-04 04:20:24 -07:00
if ! command -v ~/.acme.sh/acme.sh &> /dev/null; then
2025-12-27 15:03:33 -08:00
echo "acme.sh could not be found. Installing now..."
cd ~ || return 1
curl -s https://get.acme.sh | sh
if [ $? -ne 0 ]; then
echo -e "${red}Failed to install acme.sh${plain}"
return 1
else
echo -e "${green}acme.sh installed successfully${plain}"
fi
fi
# get the domain here, and we need to verify it
local domain=""
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ "$NONINTERACTIVE" == "1" ]]; then
domain="${XUI_DOMAIN// /}"
if [[ -z "$domain" ]] || ! is_domain "$domain"; then
echo -e "${red}XUI_SSL_MODE=domain requires a valid XUI_DOMAIN (got: '${XUI_DOMAIN:-}').${plain}"
return 1
2025-12-27 15:03:33 -08:00
fi
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
else
while true; do
read -rp "Please enter your domain name: " domain
domain="${domain// /}" # Trim whitespace
2026-05-04 04:20:24 -07:00
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ -z "$domain" ]]; then
echo -e "${red}Domain name cannot be empty. Please try again.${plain}"
continue
fi
2026-05-04 04:20:24 -07:00
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if ! is_domain "$domain"; then
echo -e "${red}Invalid domain format: ${domain}. Please enter a valid domain name.${plain}"
continue
fi
break
done
fi
2025-12-27 15:03:33 -08:00
echo -e "${green}Your domain is: ${domain}, checking it...${plain}"
SSL_ISSUED_DOMAIN="${domain}"
2025-12-27 15:03:33 -08:00
# detect existing certificate and reuse it only if its files are actually
# present and non-empty. acme.sh stores ECC certs under ${domain}_ecc and RSA
# certs under ${domain}; a failed issuance can leave a domain entry in --list
# with no usable cert files, which must not be reused (it produces a 0-byte
# fullchain.pem). Broken partial state is cleaned up so issuance can proceed.
local cert_exists=0
2026-05-04 04:20:24 -07:00
if ~/.acme.sh/acme.sh --list 2> /dev/null | awk '{print $1}' | grep -Fxq "${domain}"; then
local acmeCertDir=""
if [[ -s ~/.acme.sh/${domain}_ecc/fullchain.cer && -s ~/.acme.sh/${domain}_ecc/${domain}.key ]]; then
acmeCertDir=~/.acme.sh/${domain}_ecc
elif [[ -s ~/.acme.sh/${domain}/fullchain.cer && -s ~/.acme.sh/${domain}/${domain}.key ]]; then
acmeCertDir=~/.acme.sh/${domain}
fi
if [[ -n "${acmeCertDir}" ]]; then
cert_exists=1
local certInfo=$(~/.acme.sh/acme.sh --list 2> /dev/null | grep -F "${domain}")
echo -e "${yellow}Existing certificate found for ${domain}, will reuse it.${plain}"
[[ -n "${certInfo}" ]] && echo "$certInfo"
else
echo -e "${yellow}Found incomplete acme.sh state for ${domain} (no valid certificate files); cleaning it up and re-issuing.${plain}"
rm -rf ~/.acme.sh/${domain} ~/.acme.sh/${domain}_ecc
fi
fi
if [[ ${cert_exists} -eq 0 ]]; then
2025-12-27 15:03:33 -08:00
echo -e "${green}Your domain is ready for issuing certificates now...${plain}"
fi
# create a directory for the certificate
certPath="/root/cert/${domain}"
if [ ! -d "$certPath" ]; then
mkdir -p "$certPath"
else
rm -rf "$certPath"
mkdir -p "$certPath"
fi
# get the port number for the standalone server
local WebPort=80
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
prompt_or_default WebPort "Please choose which port to use (default is 80): " "80" XUI_ACME_HTTP_PORT
if [[ -z ${WebPort} ]]; then
WebPort=80
elif [[ ! ${WebPort} =~ ^[1-9][0-9]*$ || ${WebPort} -gt 65535 ]]; then
2025-12-27 15:03:33 -08:00
echo -e "${yellow}Your input ${WebPort} is invalid, will use default port 80.${plain}"
WebPort=80
fi
echo -e "${green}Will use port: ${WebPort} to issue certificates. Please make sure this port is open.${plain}"
# Stop panel temporarily
echo -e "${yellow}Stopping panel temporarily...${plain}"
2026-05-04 04:20:24 -07:00
systemctl stop x-ui 2> /dev/null || rc-service x-ui stop 2> /dev/null
2025-12-27 15:03:33 -08:00
if [[ ${cert_exists} -eq 0 ]]; then
# issue the certificate
~/.acme.sh/acme.sh --set-default-ca --server letsencrypt --force
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
[[ -n "${XUI_ACME_EMAIL:-}" ]] && ~/.acme.sh/acme.sh --register-account -m "${XUI_ACME_EMAIL}" > /dev/null 2>&1
~/.acme.sh/acme.sh --issue -d ${domain} $(acme_listen_flag) --standalone --httpport ${WebPort} --force
if [ $? -ne 0 ]; then
echo -e "${red}Issuing certificate failed, please check logs.${plain}"
rm -rf ~/.acme.sh/${domain} ~/.acme.sh/${domain}_ecc
2026-05-04 04:20:24 -07:00
systemctl start x-ui 2> /dev/null || rc-service x-ui start 2> /dev/null
return 1
else
echo -e "${green}Issuing certificate succeeded, installing certificates...${plain}"
fi
2025-12-27 15:03:33 -08:00
else
echo -e "${green}Using existing certificate, installing certificates...${plain}"
2025-12-27 15:03:33 -08:00
fi
# Setup reload command
reloadCmd="systemctl restart x-ui || rc-service x-ui restart"
echo -e "${green}Default --reloadcmd for ACME is: ${yellow}systemctl restart x-ui || rc-service x-ui restart${plain}"
echo -e "${green}This command will run on every certificate issue and renew.${plain}"
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ "$NONINTERACTIVE" == "1" ]]; then
setReloadcmd="n"
else
read -rp "Would you like to modify --reloadcmd for ACME? (y/n): " setReloadcmd
fi
2025-12-27 15:03:33 -08:00
if [[ "$setReloadcmd" == "y" || "$setReloadcmd" == "Y" ]]; then
echo -e "\n${green}\t1.${plain} Preset: systemctl reload nginx ; systemctl restart x-ui"
echo -e "${green}\t2.${plain} Input your own command"
echo -e "${green}\t0.${plain} Keep default reloadcmd"
read -rp "Choose an option: " choice
case "$choice" in
2026-05-04 04:20:24 -07:00
1)
echo -e "${green}Reloadcmd is: systemctl reload nginx ; systemctl restart x-ui${plain}"
reloadCmd="systemctl reload nginx ; systemctl restart x-ui"
;;
2)
echo -e "${yellow}It's recommended to put x-ui restart at the end${plain}"
read -rp "Please enter your custom reloadcmd: " reloadCmd
echo -e "${green}Reloadcmd is: ${reloadCmd}${plain}"
;;
*)
echo -e "${green}Keeping default reloadcmd${plain}"
;;
2025-12-27 15:03:33 -08:00
esac
fi
# install the certificate
local installOutput=""
installOutput=$(~/.acme.sh/acme.sh --installcert --force -d ${domain} \
2025-12-27 15:03:33 -08:00
--key-file /root/cert/${domain}/privkey.pem \
--fullchain-file /root/cert/${domain}/fullchain.pem --reloadcmd "${reloadCmd}" 2>&1)
local installRc=$?
echo "${installOutput}"
2025-12-27 15:03:33 -08:00
local installWroteFiles=0
if echo "${installOutput}" | grep -q "Installing key to:" && echo "${installOutput}" | grep -q "Installing full chain to:"; then
installWroteFiles=1
fi
2026-05-04 04:20:24 -07:00
if [[ -f "/root/cert/${domain}/privkey.pem" && -f "/root/cert/${domain}/fullchain.pem" && (${installRc} -eq 0 || ${installWroteFiles} -eq 1) ]]; then
echo -e "${green}Installing certificate succeeded, enabling auto renew...${plain}"
else
2025-12-27 15:03:33 -08:00
echo -e "${red}Installing certificate failed, exiting.${plain}"
if [[ ${cert_exists} -eq 0 ]]; then
rm -rf ~/.acme.sh/${domain} ~/.acme.sh/${domain}_ecc
fi
2026-05-04 04:20:24 -07:00
systemctl start x-ui 2> /dev/null || rc-service x-ui start 2> /dev/null
2025-12-27 15:03:33 -08:00
return 1
fi
# enable auto-renew
~/.acme.sh/acme.sh --upgrade --auto-upgrade
if [ $? -ne 0 ]; then
echo -e "${yellow}Auto renew setup had issues, certificate details:${plain}"
ls -lah /root/cert/${domain}/
2026-05-19 04:09:35 -07:00
# Secure permissions: private key readable only by owner
chmod 600 $certPath/privkey.pem 2> /dev/null
chmod 644 $certPath/fullchain.pem 2> /dev/null
2025-12-27 15:03:33 -08:00
else
echo -e "${green}Auto renew succeeded, certificate details:${plain}"
ls -lah /root/cert/${domain}/
2026-05-19 04:09:35 -07:00
# Secure permissions: private key readable only by owner
chmod 600 $certPath/privkey.pem 2> /dev/null
chmod 644 $certPath/fullchain.pem 2> /dev/null
2025-12-27 15:03:33 -08:00
fi
2026-05-19 04:09:35 -07:00
# start panel
2026-05-04 04:20:24 -07:00
systemctl start x-ui 2> /dev/null || rc-service x-ui start 2> /dev/null
2025-12-27 15:03:33 -08:00
# Prompt user to set panel paths after successful certificate installation
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ "$NONINTERACTIVE" == "1" ]]; then
setPanel="y"
else
read -rp "Would you like to set this certificate for the panel? (y/n): " setPanel
fi
2025-12-27 15:03:33 -08:00
if [[ "$setPanel" == "y" || "$setPanel" == "Y" ]]; then
local webCertFile="/root/cert/${domain}/fullchain.pem"
local webKeyFile="/root/cert/${domain}/privkey.pem"
if [[ -f "$webCertFile" && -f "$webKeyFile" ]]; then
${xui_folder}/x-ui cert -webCert "$webCertFile" -webCertKey "$webKeyFile"
2025-12-27 15:03:33 -08:00
echo -e "${green}Certificate paths set for the panel${plain}"
echo -e "${green}Certificate File: $webCertFile${plain}"
echo -e "${green}Private Key File: $webKeyFile${plain}"
echo ""
echo -e "${green}Access URL: https://${domain}:${existing_port}/${existing_webBasePath}${plain}"
echo -e "${yellow}Panel will restart to apply SSL certificate...${plain}"
2026-05-04 04:20:24 -07:00
systemctl restart x-ui 2> /dev/null || rc-service x-ui restart 2> /dev/null
2025-12-27 15:03:33 -08:00
else
echo -e "${red}Error: Certificate or private key file not found for domain: $domain.${plain}"
fi
else
echo -e "${yellow}Skipping panel path setting.${plain}"
fi
2026-05-04 04:20:24 -07:00
2025-12-27 15:03:33 -08:00
return 0
}
2026-05-19 04:09:35 -07:00
# Reusable interactive SSL setup (domain or IP)
# Sets global `SSL_HOST` to the chosen domain/IP for Access URL usage
2025-12-27 15:03:33 -08:00
prompt_and_setup_ssl() {
local panel_port="$1"
2026-05-19 04:09:35 -07:00
local web_base_path="$2"
2025-12-27 15:03:33 -08:00
local server_ip="$3"
local ssl_choice=""
2026-05-19 04:09:35 -07:00
SSL_SCHEME="https"
2025-12-27 15:03:33 -08:00
echo -e "${yellow}Choose SSL certificate setup method:${plain}"
2026-01-04 20:28:02 -08:00
echo -e "${green}1.${plain} Let's Encrypt for Domain (90-day validity, auto-renews)"
echo -e "${green}2.${plain} Let's Encrypt for IP Address (6-day validity, auto-renews)"
echo -e "${green}3.${plain} Custom SSL Certificate (Path to existing files)"
2026-05-19 04:09:35 -07:00
echo -e "${green}4.${plain} Skip SSL (advanced — behind reverse proxy / SSH tunnel only)"
echo -e "${blue}Note:${plain} Options 1 & 2 require port 80 open. Option 3 requires manual paths."
2026-05-19 04:09:35 -07:00
echo -e "${blue}Note:${plain} Option 4 serves the panel over plain HTTP — only safe behind nginx/Caddy or an SSH tunnel."
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ "$NONINTERACTIVE" == "1" ]]; then
case "${XUI_SSL_MODE:-none}" in
domain) ssl_choice="1" ;;
ip) ssl_choice="2" ;;
none | "") ssl_choice="4" ;;
*)
echo -e "${yellow}Unknown XUI_SSL_MODE='${XUI_SSL_MODE}', defaulting to none (HTTP).${plain}"
ssl_choice="4"
;;
esac
else
read -rp "Choose an option (default 2 for IP): " ssl_choice
ssl_choice="${ssl_choice// /}" # Trim whitespace
2026-05-04 04:20:24 -07:00
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
# Default to 2 (IP cert) if input is empty or invalid (not 1, 3 or 4)
if [[ "$ssl_choice" != "1" && "$ssl_choice" != "3" && "$ssl_choice" != "4" ]]; then
ssl_choice="2"
fi
2025-12-27 15:03:33 -08:00
fi
case "$ssl_choice" in
2026-05-04 04:20:24 -07:00
1)
# User chose Let's Encrypt domain option
echo -e "${green}Using Let's Encrypt for domain certificate...${plain}"
if ssl_cert_issue; then
local cert_domain="${SSL_ISSUED_DOMAIN}"
if [[ -z "${cert_domain}" ]]; then
cert_domain=$(~/.acme.sh/acme.sh --list 2> /dev/null | tail -1 | awk '{print $1}')
fi
2026-05-04 04:20:24 -07:00
if [[ -n "${cert_domain}" ]]; then
SSL_HOST="${cert_domain}"
echo -e "${green}✓ SSL certificate configured successfully with domain: ${cert_domain}${plain}"
else
echo -e "${yellow}SSL setup may have completed, but domain extraction failed${plain}"
SSL_HOST="${server_ip}"
fi
else
2026-05-04 04:20:24 -07:00
echo -e "${red}SSL certificate setup failed for domain mode.${plain}"
SSL_HOST="${server_ip}"
fi
2026-05-04 04:20:24 -07:00
;;
2)
# User chose Let's Encrypt IP certificate option
echo -e "${green}Using Let's Encrypt for IP certificate (shortlived profile)...${plain}"
# Confirm the auto-detected IP before issuing for it: with asymmetric
# routing / multi-WAN the echo services can return a transit address.
if [[ "$NONINTERACTIVE" != "1" ]]; then
local ip_confirm=""
read -rp "Is ${server_ip} the correct incoming public IPv4 address for this server? [Default y]: " ip_confirm
if [[ -n "$ip_confirm" && "$ip_confirm" != "y" && "$ip_confirm" != "Y" ]]; then
server_ip=""
while [[ -z "$server_ip" ]]; do
read -rp "Please enter your server's public IPv4 address: " server_ip
server_ip="${server_ip// /}"
if [[ ! "$server_ip" =~ ^[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
echo -e "${red}Invalid IPv4 address. Please try again.${plain}"
server_ip=""
fi
done
fi
fi
2026-05-04 04:20:24 -07:00
# Ask for optional IPv6
local ipv6_addr=""
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
prompt_or_default ipv6_addr "Do you have an IPv6 address to include? (leave empty to skip): " "" XUI_SSL_IPV6
2026-05-04 04:20:24 -07:00
ipv6_addr="${ipv6_addr// /}" # Trim whitespace
2026-05-04 04:20:24 -07:00
# Stop panel if running (port 80 needed)
if [[ $release == "alpine" ]]; then
rc-service x-ui stop > /dev/null 2>&1
else
2026-05-04 04:20:24 -07:00
systemctl stop x-ui > /dev/null 2>&1
fi
2026-05-04 04:20:24 -07:00
setup_ip_certificate "${server_ip}" "${ipv6_addr}"
if [ $? -eq 0 ]; then
SSL_HOST="${server_ip}"
echo -e "${green}✓ Let's Encrypt IP certificate configured successfully${plain}"
else
2026-05-04 04:20:24 -07:00
echo -e "${red}✗ IP certificate setup failed. Please check port 80 is open.${plain}"
SSL_HOST="${server_ip}"
fi
2026-05-04 04:20:24 -07:00
;;
3)
# User chose Custom Paths (User Provided) option
echo -e "${green}Using custom existing certificate...${plain}"
local custom_cert=""
local custom_key=""
local custom_domain=""
2026-05-04 04:20:24 -07:00
# 3.1 Request Domain to compose Panel URL later
read -rp "Please enter domain name certificate issued for: " custom_domain
custom_domain="${custom_domain// /}" # Remove spaces
# 3.2 Loop for Certificate Path
while true; do
read -rp "Input certificate path (keywords: .crt / fullchain): " custom_cert
# Strip quotes if present
custom_cert=$(echo "$custom_cert" | tr -d '"' | tr -d "'")
if [[ -f "$custom_cert" && -r "$custom_cert" && -s "$custom_cert" ]]; then
break
elif [[ ! -f "$custom_cert" ]]; then
echo -e "${red}Error: File does not exist! Try again.${plain}"
elif [[ ! -r "$custom_cert" ]]; then
echo -e "${red}Error: File exists but is not readable (check permissions)!${plain}"
else
echo -e "${red}Error: File is empty!${plain}"
fi
done
# 3.3 Loop for Private Key Path
while true; do
read -rp "Input private key path (keywords: .key / privatekey): " custom_key
# Strip quotes if present
custom_key=$(echo "$custom_key" | tr -d '"' | tr -d "'")
if [[ -f "$custom_key" && -r "$custom_key" && -s "$custom_key" ]]; then
break
elif [[ ! -f "$custom_key" ]]; then
echo -e "${red}Error: File does not exist! Try again.${plain}"
elif [[ ! -r "$custom_key" ]]; then
echo -e "${red}Error: File exists but is not readable (check permissions)!${plain}"
else
echo -e "${red}Error: File is empty!${plain}"
fi
done
# 3.4 Apply Settings via x-ui binary
${xui_folder}/x-ui cert -webCert "$custom_cert" -webCertKey "$custom_key" > /dev/null 2>&1
# Set SSL_HOST for composing Panel URL
if [[ -n "$custom_domain" ]]; then
SSL_HOST="$custom_domain"
else
SSL_HOST="${server_ip}"
fi
echo -e "${green}✓ Custom certificate paths applied.${plain}"
echo -e "${yellow}Note: You are responsible for renewing these files externally.${plain}"
systemctl restart x-ui > /dev/null 2>&1 || rc-service x-ui restart > /dev/null 2>&1
;;
2026-05-19 04:09:35 -07:00
4)
echo ""
echo -e "${red}⚠ Panel will be installed WITHOUT SSL/TLS.${plain}"
echo -e "${yellow}Login credentials and cookies will travel as plain HTTP.${plain}"
echo -e "${yellow}Only safe when:${plain}"
echo -e "${yellow} • A reverse proxy (nginx, Caddy, Traefik) terminates TLS for you, or${plain}"
echo -e "${yellow} • You access the panel exclusively via SSH tunnel${plain}"
echo ""
SSL_SCHEME="http"
SSL_HOST="${server_ip}"
local bind_local=""
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ "$NONINTERACTIVE" == "1" ]]; then
# Cloud images must stay reachable on their public interface.
bind_local="n"
else
read -rp "Bind the panel to 127.0.0.1 only? (recommended — forces SSH tunnel / reverse-proxy access) [y/N]: " bind_local
fi
2026-05-19 04:09:35 -07:00
if [[ "$bind_local" == "y" || "$bind_local" == "Y" ]]; then
${xui_folder}/x-ui setting -listenIP "127.0.0.1" > /dev/null 2>&1
SSL_HOST="127.0.0.1"
echo -e "${green}✓ Panel bound to 127.0.0.1 only. It is now unreachable from the public internet.${plain}"
echo ""
echo -e "${green}SSH Port Forwarding — open the panel from your local machine via:${plain}"
echo -e " Standard SSH command:"
echo -e " ${yellow}ssh -L 2222:127.0.0.1:${panel_port} root@${server_ip}${plain}"
echo -e " If using an SSH key:"
echo -e " ${yellow}ssh -i <sshkeypath> -L 2222:127.0.0.1:${panel_port} root@${server_ip}${plain}"
echo -e " Then open in your browser:"
echo -e " ${yellow}http://localhost:2222/${web_base_path}${plain}"
echo ""
echo -e "${yellow}Alternative: point a reverse proxy (nginx/Caddy) at 127.0.0.1:${panel_port} and let it terminate TLS.${plain}"
else
echo -e "${yellow}Panel will listen on all interfaces over plain HTTP. Make sure something else is terminating TLS in front of it.${plain}"
fi
systemctl restart x-ui > /dev/null 2>&1 || rc-service x-ui restart > /dev/null 2>&1
echo -e "${green}✓ SSL setup skipped.${plain}"
;;
2026-05-04 04:20:24 -07:00
*)
echo -e "${red}Invalid option. Skipping SSL setup.${plain}"
SSL_HOST="${server_ip}"
2026-05-04 04:20:24 -07:00
;;
2025-12-27 15:03:33 -08:00
esac
}
2026-05-19 04:09:35 -07:00
config_after_install() {
local existing_hasDefaultCredential=$(${xui_folder}/x-ui setting -show true | grep -Eo 'hasDefaultCredential: .+' | awk '{print $2}')
Feat/multi inbound clients (#4469) * feat(clients): add shadow tables for first-class client promotion Introduces three new GORM-backed tables (clients, client_inbounds, inbound_fallback_children) and a populate-only seeder that backfills them from each inbound's existing settings.clients JSON. Duplicate emails across inbounds auto-merge under one client row, with each field conflict logged. Existing services are unchanged and continue reading from settings.clients — this commit is groundwork only. * feat(clients): make clients+client_inbounds the runtime source of truth Adds ClientService.SyncInbound that reconciles the new tables from each inbound's clients list whenever existing service paths mutate settings.clients. Wires it into AddInbound, UpdateInbound, AddInboundClient, UpdateInboundClient, DelInboundClient, DelInboundClientByEmail, DelDepletedClients, autoRenewClients, and the timestamp-backfill path in adjustTraffics, plus DetachInbound on DelInbound. GetXrayConfig now builds settings.clients from the new tables before writing config.json, and getInboundsBySubId joins through them instead of JSON_EACH on settings JSON. Live Xray config and subscription endpoints are now driven by the relational view; settings.clients JSON stays in step as a side effect of every write. * feat(clients): add top-level Clients tab and CRUD API Adds /panel/api/clients endpoints (list, get, add, update, del, attach, detach) backed by ClientService methods that orchestrate the per-inbound Add/Update/Del flows so a single client row is created once and attached to many inbounds in one operation. The frontend gains a dedicated Clients page (frontend/clients.html + src/pages/clients/) with an AntD table, multi-inbound attach modal, and full CRUD. Axios interceptor learns to honour Content-Type: application/json so the JSON endpoints work alongside the legacy form-encoded ones. The legacy per-inbound client modal stays untouched in this PR — both flows now write to the same source of truth. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds): add Port-with-Fallback inbound type Adds a new "portfallback" protocol that emits as a VLESS-TLS inbound under the hood but is paired with a sidecar table of child inbounds. Panel auto-builds settings.fallbacks at Xray-config-gen time from the sidecar — each child's listen+port becomes the fallback dest, with SNI/ALPN/path/xver match criteria pulled from the row. No more typing loopback ports by hand or keeping settings.fallbacks in sync. Backend: new FallbackService (Get/SetChildren, BuildFallbacksJSON); two new routes (GET/POST /panel/api/inbounds/:id/fallbackChildren); xray.GetXrayConfig injects fallbacks for PortFallback inbounds; the inbound model emits protocol="vless" so Xray accepts the config. Frontend: PORTFALLBACK joins the protocol dropdown; selecting it shows the standard VLESS controls plus a Fallback Children table (inbound picker + per-row SNI/ALPN/path/xver). Children are loaded on edit and replaced atomically on save. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Reset Traffic, QR Code, Info actions + Online/Remaining columns The Clients page table gains: - Online column — green/grey tag driven by /panel/api/inbounds/onlines, polled every 10s. - Remaining column — bytes-remaining tag, coloured green/orange/red against quota, purple infinity when unlimited. - Action icons per row: QR, Info, Reset traffic, Edit, Delete. ClientInfoModal shows the full client detail (uuid/password/auth, traffic ↑/↓ + remaining + all-time, expiry absolute + relative, attached inbounds chip list, online + last-online). ClientQrModal fetches links for the client's subId via /panel/api/inbounds/getSubLinks/:subId and renders each one through the existing QrPanel component. Reset Traffic confirms then calls the existing per-inbound endpoint on the client's first attached inbound (the traffic row is keyed on email globally, so any attached inbound resets the shared counter). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): expose Attached inbounds in edit mode The multi-select was gated on add-only, so editing a client had no way to change which inbounds it belonged to. The picker now shows in both modes, and on submit the modal diffs the picked set against the original attachedIds — additions go through the /attach endpoint, removals through /detach, both after the field update lands so the new attachments get the latest values. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): unbreak template parsing + stale i18n keys - InboundFormModal: split the multi-line help string in the PortFallback section onto one line — Vue's template parser was bailing on Unterminated string constant because a single-quoted literal spanned two lines inside a {{ }} interpolation. - ClientInfoModal: t('disable') was missing at the root level, so vue-i18n returned the key path literally. Use t('disabled') which exists. - Linter cleanup elsewhere: pages.client.* references renamed to pages.clients.* to match the merged i18n block; whitespace normalisation in a few unrelated Vue templates. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * 1 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(traffic): drop all-time traffic tracking Removes the AllTime field from Inbound and ClientTraffic and migrates existing DBs by dropping the all_time columns on startup. The counter duplicated up+down without adding signal, and the per-event accumulator ran on every traffic write. Frontend: drop the All-time column from the inbound list and the client-row table, the All-time row from the client info modal, and the All-Time Total Usage tile from the inbounds summary card. The allTimeTraffic/allTimeTrafficUsage i18n keys are removed across every locale. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): mobile cards, multi-select, bulk add Adds the same row-card layout the inbounds page uses on mobile: the table is suppressed under the mobile breakpoint and each client renders as a compact card with a status dot, email, Info button, Enable switch, and overflow menu. All the per-client detail (traffic, remaining, expiry, attached inbounds, flow, created/updated, URL, subscription) opens through the existing info modal. Multi-select with bulk delete wires AntD row-selection on desktop and a per-card checkbox on mobile; a Delete (N) button appears in the toolbar when anything is selected. Bulk add reuses the five email-generation modes from the inbound bulk modal but takes a multi-inbound picker so one bulk run can attach to several inbounds at once. Submits client-by-client through the existing /panel/api/clients/add endpoint. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(inbounds): remove legacy per-inbound client UI Now that clients live as first-class rows attached to one or many inbounds, the per-inbound client UI on the inbounds page is dead weight — every client action either has a global equivalent on the Clients page or makes no sense in a many-to-many world. Deletes ClientFormModal, ClientBulkModal, CopyClientsModal, and ClientRowTable from inbounds/. Strips the matching emits, refs, handlers, and dropdown menu items from InboundList and InboundsPage, and removes the dead mobile expand-chevron state and the desktop expanded-row plumbing that drove the inline client table. The InboundFormModal Clients tab still works in add-mode (one inline client at inbound creation) — that flow goes through ClientService. SyncInbound on save and remains useful. Fixes a stray "</a-dropdown>" left over by an earlier toolbar edit in ClientsPage that broke the template parser. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Delete depleted action Mirrors the legacy delDepletedClients action that lived under the inbounds page, but as a first-class /panel/api/clients/delDepleted endpoint backed by ClientService. The new path goes through ClientService.Delete for each depleted email, so the new clients + client_inbounds + xray_client_traffic tables stay consistent. Adds a danger-styled toolbar button on the Clients page (next to Reset all client traffic) with a confirm dialog and a toast reporting the deleted count. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(api): move every client-shaped endpoint off /inbounds onto /clients After the multi-inbound client migration, client state belongs to the client API surface, not the inbound one. Twelve routes that were crammed under /panel/api/inbounds/* now live where they belong, under /panel/api/clients/*. Moved (route, handler, doc): POST /clientIps/:email POST /clearClientIps/:email POST /onlines POST /lastOnline POST /updateClientTraffic/:email POST /resetAllClientTraffics/:id POST /delDepletedClients/:id POST /:id/resetClientTraffic/:email GET /getClientTraffics/:email GET /getClientTrafficsById/:id GET /getSubLinks/:subId GET /getClientLinks/:id/:email Their /clients/* counterparts are: POST /clients/clientIps/:email POST /clients/clearClientIps/:email POST /clients/onlines POST /clients/lastOnline POST /clients/updateTraffic/:email POST /clients/resetTraffic/:email (email-only, fans out) GET /clients/traffic/:email GET /clients/traffic/byId/:id GET /clients/subLinks/:subId GET /clients/links/:id/:email per-inbound resetAllClientTraffics and delDepletedClients are dropped entirely — the Clients page already exposes global Reset All Traffic and Delete depleted actions, and per-inbound resets are meaningless once a client can be attached to many inbounds. ClientService.ResetTrafficByEmail is the new email-only reset path: it looks up every inbound the client is attached to and pushes the counter reset + Xray re-add through inboundService.ResetClientTraffic for each one, so depleted users come back online instantly. Frontend callers (ClientsPage, useClients, ClientQrModal, ClientInfoModal, InboundInfoModal, InboundsPage, useInbounds) all switched to the new paths. The Inbounds page drops its per-inbound "Reset client traffic" and "Delete depleted clients" dropdown items — users do those at the client level now. api-docs is rebuilt to match. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(service): switch tgbot + ldap callers to ClientService Adds two thin helpers to ClientService (CreateOne, DetachByEmail) and rewrites tgbot.SubmitAddClient and ldap_sync_job to call ClientService directly. Removes the JSON-blob payloads (BuildJSONForProtocol output for add, clientsToJSON/clientToJSON helpers) that callers previously fed to InboundService.AddInboundClient/DelInboundClient. ldap_sync_job.batchSetEnable now loops InboundService.SetClientEnableByEmail per email instead of trying to coerce AddInboundClient into doing the update — the old path would have failed duplicate-email validation for existing clients anyway. The legacy InboundService.AddInboundClient/UpdateInboundClient/ DelInboundClient methods stay in place; they are now only used internally by ClientService Create/Update/Delete/Attach. Inlining + deleting them follows in a separate commit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(service): move all client mutation methods to ClientService Moves the client mutation surface out of InboundService and into ClientService. These methods all operate on a single client (identity fields, traffic limits, expiry, ip limit, enable state, telegram tg id) and didn't belong on the inbound aggregate. Moved (12 methods): AddInboundClient, UpdateInboundClient, DelInboundClient, DelInboundClientByEmail, checkEmailsExistForClients, SetClientTelegramUserID, checkIsEnabledByEmail, ToggleClientEnableByEmail, SetClientEnableByEmail, ResetClientIpLimitByEmail, ResetClientExpiryTimeByEmail, ResetClientTrafficLimitByEmail. Each method now takes an explicit *InboundService for the helpers that legitimately stay on InboundService (GetInbound, GetClients, runtimeFor, AddClientStat / UpdateClientStat / DelClientStat, DelClientIPs / UpdateClientIPs, emailUsedByOtherInbounds, getAllEmailSubIDs, GetClientInboundByEmail / GetClientInboundByTrafficID, GetClientTrafficByEmail). Stays on InboundService: ResetClientTrafficByEmail and ResetClientTraffic(id, email) — these mutate xray_client_traffic rows, not client identity, so they're inbound-side bookkeeping. Callers updated: tgbot (6 calls), ldap_sync_job (1 call), InboundService internal (writeBackClientSubID, CopyInboundClients, AddInbound's email-uniqueness check), ClientService Create/Update/ Delete/Attach/Detach. Also removes a dead resetAllClientTraffics controller handler whose route was already gone after the previous /clients API migration. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(clients): finish migrating to ClientService + tidy IP routes Two related cleanups in the new /clients surface: 1. Move ResetAllClientTraffics (bulk-reset of xray_client_traffic + last_traffic_reset_time, with node-runtime propagation) from InboundService to ClientService. PeriodicTrafficResetJob now holds a clientService and calls j.clientService.ResetAllClientTraffics(&j.inboundService, id). The last client-mutation method on InboundService is gone. 2. Shorten redundantly-named routes/handlers under /panel/api/clients: - /clientIps/:email -> /ips/:email (handler getIps) - /clearClientIps/:email -> /clearIps/:email (handler clearIps) The "client" prefix was redundant inside the clients namespace. Frontend (InboundInfoModal) and api-docs updated to match. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds,clients): clean up inbound modal + enrich client modal Inbound modal rework (InboundFormModal.vue + inbound.js): - Drop the embedded Client subform in the Protocol tab. Multi-inbound clients are managed exclusively from the Clients page now; a fresh inbound is created with zero clients (settings constructors default to []) and the user attaches clients afterwards. - Hide the Protocol tab entirely when it has nothing to render (VMESS, Trojan without fallbacks, Hysteria). Auto-switches active tab to Basic when the tab disappears while focused. - Move the Security section (Security selector + TLS block with certs and ECH + Reality block) out of the Stream tab into its own Security tab, sharing the canEnableStream gate. Client modal additions (ClientFormModal.vue + ClientBulkAddModal.vue): - Flow select (xtls-rprx-vision / -udp443) appears only when the panel actually has a Vision-capable inbound (VLESS or PortFallback on TCP with TLS or Reality). Hidden otherwise, and cleared when it disappears. - IP Limit input is disabled when the panel-level ipLimitEnable setting is off, fetched into useClients alongside subSettings and threaded through ClientsPage to both modals. - Edit modal now shows an "IP Log" section listing IPs that have connected with the client's credentials, with refresh and clear buttons (calls the renamed /panel/api/clients/ips and /clearIps endpoints). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(inbounds): drop manual Fallbacks UI from inbound modal The PortFallback protocol type now covers the common VLESS-master-plus-children case with auto-wired dests, so the manual Fallbacks editor (showFallbacks block in the Protocol tab) is mostly redundant. Removed: - the v-if="showFallbacks" template block (SNI/ALPN/Path/dest/PROXY rows) - the showFallbacks computed - the addFallback / delFallback helpers - the .fallbacks-header / .fallbacks-title styles - the showFallbacks gate from hasProtocolTabContent (so Trojan-over-TCP no longer shows an empty Protocol tab) Power users who need a non-inbound fallback dest (nginx, static site) can still author settings.fallbacks via the Advanced JSON tab. * feat(clients,inbounds): move search/filter to Clients page + small fixes Search/filter relocation: - Remove the search/filter toolbar (search switch + filter radio + protocol/node selects + the visibleInbounds projection + inboundsFilterState localStorage + filter CSS + the SearchOutlined/ FilterOutlined/ObjectUtil/Inbound imports it required) from InboundList. The filters were all client-oriented buckets bolted onto the inbound row. - Add a search/filter toolbar to ClientsPage with the same shape: switch between deep-text search and bucket filter (active / deactive / depleted / expiring / online) + protocol filter that matches clients attached to at least one inbound with the chosen protocol. State persists in clientsFilterState localStorage. filteredClients drives both the desktop table and the mobile card list, and select-all / allSelected / someSelected only span the visible subset. - useClients now also fetches expireDiff and trafficDiff from /panel/setting/defaultSettings (used to detect the expiring bucket); ClientsPage threads them into the client-bucket helper. Loose fixes folded in: - Add Client: email field is auto-filled with a random handle on open, matching uuid/subId/password/auth. - Inbound clone: parse and reuse the source settings JSON (with clients reset to []) instead of building a fresh defaulted Settings, so VLESS Encryption/Decryption and other non-client fields survive the clone. - en-US.json: add the ipLog string used by the edit-client modal. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Reverse tag field for VLESS-attached clients Mirrors the Flow field's pattern: a Reverse tag input appears in the Add/Edit Client modal whenever at least one selected inbound is VLESS or PortFallback. The value rides over the wire as client.reverse = { tag: '...' } so it lands directly in model.Client's *ClientReverse field; an empty value omits the reverse key entirely. On edit the field is hydrated from props.client.reverse?.tag, and the showReverseTag watcher clears the field if the user drops the last VLESS-like inbound from the selection. * fix(xray): emit only protocol-relevant fields per client entry The Xray config synthesizer was writing every identifier field (id, password, flow, auth, security/method, reverse) on every client entry regardless of the inbound's protocol. Xray ignores unknown fields, so the config worked, but it diverged from the spec and leaked secrets across protocols when one client was attached to multiple inbounds — a VLESS inbound's generated config carried the same client's Trojan password and Hysteria auth alongside its uuid. Switch on inbound.Protocol when building each entry: - VLESS / PortFallback: id, flow, reverse - VMess: id, security - Trojan: password, flow - Shadowsocks: password, method - Hysteria / Hysteria2: auth email is emitted for every protocol. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): restore auto-disable kick under new schema disableInvalidClients still resolved (inbound_tag, email) pairs via JSON_EACH(inbounds.settings.clients), which is empty after migrating to the clients + client_inbounds tables. Result: xrayApi.RemoveUser never ran for depleted clients, clients.enable stayed true so the UI showed them as active, and only xray_client_traffic.enable got flipped - making "Restart Xray After Auto Disable" only half-work. Resolve the targets via a JOIN through the new schema, flip clients.enable so the Clients page reflects the state, and drop the legacy JSON write-back plus the subId cascade workaround (email is unique now). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): live WebSocket updates + Ended status surfacing ClientsPage now subscribes to traffic / client_stats / invalidate WebSocket events instead of polling /onlines every 10s. Per-row traffic counters refresh in place, online state stays current, and list-level mutations elsewhere trigger a refresh. The client roll-up summary moves from InboundsPage to ClientsPage where it belongs, restructured into six labeled stat tiles (Total / Online / Ended / Expiring / Disabled / Active) with email popovers on the ones with issues. Auto-disabled clients (traffic exhausted or expiry passed) now classify as 'depleted' even though clients.enable=false, so they show up under the Ended filter and render a red Ended tag instead of looking indistinguishable from an operator-disabled row. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(nodes): per-node client roll-up and panel version Added transient inboundCount / clientCount / onlineCount / depletedCount fields to model.Node, populated by NodeService.GetAll via aggregated queries (one join across inbounds + client_inbounds, one over client_traffics intersected with the in-memory online emails). The Nodes list renders these as colored chips on a new "Clients" column so an operator can see at a glance how many users each node carries and how many are currently online or depleted. Also exposes the remote panel's version. The central panel adds panelVersion to its /api/server/status payload (sourced from config.GetVersion). Probe reads that field and persists it on the node row, mirroring how xrayVersion already flows. NodesPage gets a new column next to Xray Version, in both desktop and mobile views, with English and Persian strings. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): stop node sync from resurrecting deleted clients Several related issues around node-managed clients: - Remote runtime: drop the per-inbound resetAllClientTraffics path and point traffic/onlines/lastOnline fetches at the new /panel/api/clients/* routes. - Delete from master: always push the updated inbound to the node even when the client was already disabled or depleted, so the node actually loses the user instead of silently keeping it. - setRemoteTraffic: mirror remote clients into the central tables only on first discovery of a node inbound. Matched inbounds let the master own the join table, so a stale snap can no longer re-create a ClientRecord (and join row) for a client that was just deleted on the master. - ClientService.Delete: route through submitTrafficWrite so deletes serialize with node traffic merges, and switch the final ClientRecord delete to an explicit Where("id = ?") clause. - setRemoteTraffic UNIQUE-constraint fix: use clause.OnConflict on inserts and email-keyed UPDATEs for client_traffics, so mirroring a snap doesn't trip the unique email index. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(clients): switch client API endpoints from id to email All client-scoped routes now use the unique email as the path key (get, update, del, attach, detach, links). Email is the stable, protocol-independent identifier — UUIDs don't exist for trojan or shadowsocks, and internal numeric ids leaked panel implementation detail into the public API. Removed the redundant /traffic/byId/:id endpoint (covered by /traffic/:email) and collapsed /links/:id/:email into /links/:email, which now returns links across every attached inbound for the client. Frontend selection, bulk delete, and toggle state are now keyed by email as well, dropping the id→email lookup workaround. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(server): move cached state and helpers into ServerService ServerController had grown to hold its own status cache, version-list TTL cache, history-bucket whitelist, and the loop that drove all three — concerns that belong in the service layer. Pull them out: - lastStatus + the @2s refresh become ServerService.RefreshStatus and ServerService.LastStatus; the controller's cron now just orchestrates the cross-service side effects (xrayMetrics sample, websocket broadcast). - The 15-minute Xray-versions cache (with stale-on-error fallback) moves into ServerService.GetXrayVersionsCached, collapsing the controller handler to a single call. - The freedom/blackhole outbound-tag walk used by /xraylogs becomes ServerService.GetDefaultLogOutboundTags. - The allowed-history-bucket whitelist moves to package-level service.IsAllowedHistoryBucket, so both NodeController and ServerController validate against the same list. Net result: web/controller/server.go drops from 458 to 365 lines and contains only HTTP wiring + presentation-y side effects. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(api): emit JSON-text columns as nested objects Inbound, ClientRecord, and InboundClientIps store settings / streamSettings / sniffing / reverse / ips as JSON-text in the DB. The API was passing that text through verbatim, so every consumer had to JSON.parse a string inside a string. Add MarshalJSON / UnmarshalJSON so the wire format is a real nested object, while still accepting the legacy escaped-string shape on write. Frontend dbinbound.js gets a matching coerceInboundJsonField helper for the same dual-shape read path, and inbound.js toJson stops emitting empty/placeholder fields (externalProxy [], sniffing destOverride when disabled, etc.) so the new normalised JSON stays terse. api-docs and the inbound-clone path are updated to the new shape. Controller route lists are regrouped so all GETs sit above POSTs. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): include inboundIds and traffic in /clients/list ClientRecord got its own MarshalJSON in the previous commit, and ClientWithAttachments embeds it to add inboundIds and traffic. Go promotes the embedded MarshalJSON to the outer struct, so the encoder was calling ClientRecord.MarshalJSON for the whole value and silently dropping the extras. The frontend reads row.inboundIds / row.traffic from /clients/list, so attached inbounds didn't render and newly added clients looked like they hadn't saved. Add an explicit MarshalJSON on ClientWithAttachments that splices the extras in. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): gate IP Log on ipLimitEnable + clean access-log dropdown Legacy panel hid the IP Log section when access logging was off; the Vue 3 migration left it gated on isEdit only, so the section showed even when xray's access log was 'none' and nothing was being recorded. Restore the ipLimitEnable gate on the edit modal's IP Log form-item. While here, clean up the Xray Settings access-log dropdown: previously two 'none' entries appeared (an empty value labelled with t('none') and the literal 'none' from the options array). Drop the empty option for access log (the literal 'none' covers it) and relabel the empty option for error log / mask address to t('empty') so they're distinguishable. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(nodes): route per-client ops through node clients API + orphan sweep Adds Runtime methods AddClient, UpdateUser, and DeleteUser so master mutates clients on a node via /panel/api/clients/{add,update,del} rather than pushing the whole inbound. The previous rt.UpdateInbound path made the node DelInbound+AddInbound on every single-client change, briefly cycling every other user on the same inbound. DelInbound no longer filters by enable=true, so a disabled node inbound actually gets removed from the node instead of being resurrected by the next snap. setRemoteTrafficLocked now sweeps any ClientRecord with zero ClientInbound rows after SyncInbound rebuilds the attachments, which is how a node-side delete propagates back to master instead of leaving a detached ghost. ClientService.Delete tombstones the email first so a snap arriving mid-delete can't re-create the record. WebSocket broadcasts an "invalidate(clients)" message on every client mutation so the Clients page refreshes without manual reload. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(balancers): allow fallback on all strategies + feed burstObservatory from random/roundRobin Drops the random/roundRobin gate on the Fallback field in BalancerFormModal so every strategy can pick a fallback outbound. syncObservatories now feeds burstObservatory from leastLoad + random + roundRobin balancers (was leastLoad only), matching how leastPing feeds observatory. Fix the JsonEditor "Unexpected end of JSON input" that appeared when switching a balancer between leastPing and another strategy: the obsView watcher was gated on showObsEditor (a boolean OR of the two flags) and missed the case where one observatory swapped for the other in the same tick. Watch the individual flags instead so obsView flips to the surviving editor and the getter stops pointing at a deleted key. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(inbounds): use sortedInbounds for mobile empty-state check InboundList referenced an undefined visibleInbounds in the mobile card list's empty-state guard, throwing "Cannot read properties of undefined (reading 'length')" and breaking the entire mobile render. * feat(clients): sortable table columns Adds the same sortState / sortableCol / sortFns pattern InboundList uses, wrapping filteredClients in sortedClients so sort composes with the existing search/filter pipeline. Sortable: enable, email, inboundIds (attachment count), traffic, remaining, expiryTime; actions and online stay unsorted. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(shadowsocks): generate valid ss2022 keys and per-client method for legacy ciphers The Add Client flow on shadowsocks inbounds was producing xray configs that failed to start: - 2022-blake3-* ciphers need a base64-encoded key of an exact byte length per cipher. fillProtocolDefaults was assigning a uuid-style string, which xray rejects as "bad key". Now the password is generated (or replaced if invalid) via random.Base64Bytes(n) sized to the chosen cipher. - Legacy ciphers (aes-256-gcm, chacha20-*, xchacha20-*) require a per-client method field in multi-user mode; model.Client has no Method, so settings.clients was stored without one and xray failed with "unsupported cipher method:". applyShadowsocksClientMethod now injects the top-level method into each client on add/update, and healShadowsocksClientMethods backfills it at xray-config-build time so existing inbounds heal on the next start. - xray/api.go ssCipherType switch was missing aes-256-gcm, which fell through to ss2022 path. - SSMethods dropdown now offers aes-256-gcm. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): preserve ClientRecord on inbound delete + filter Attached inbounds to multi-client protocols Replace the global orphan sweep in setRemoteTrafficLocked with a per-inbound diff cleanup: only delete a ClientRecord whose email disappeared from a snap-tracked inbound (i.e. a node-side delete). Inbounds that vanished entirely from the snap (e.g. admin deleted the inbound on master) aren't iterated, so a client whose last attachment came from that inbound is now left alone instead of being deleted alongside the inbound. ClientFormModal and ClientBulkAddModal now filter the Attached inbounds dropdown to protocols that actually support multiple clients: shadowsocks, vless, vmess, trojan, hysteria, hysteria2, and portfallback (which routes through VLESS settings). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): make empty-state text readable on dark/ultra themes The "No clients yet" empty state had a hardcoded black color (rgba(0,0,0,0.45)) that vanished against the dark backgrounds. Drop the inline color, let it inherit from the AntD theme, and fade with opacity like the mobile card empty state already does. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): client-first tgbot add flow, tgId field, lightweight inbound options - tgbot: drop legacy per-protocol Add Client UI in favour of a client-first multi-inbound flow. New BuildClientDraftMessage / getInboundsAttachPicker let an admin pick one or more inbounds and submit a single client; per- protocol secrets are now generated server-side via fillProtocolDefaults. Drops awaiting_id/awaiting_password_tr/awaiting_password_sh state cases and add_client_ch_default_id/pass_tr/pass_sh/flow callbacks. Adds a setTGUser button + awaiting_tg_id state so the bot can set Client.TgID during Add. - clients UI: add Telegram user ID input to ClientFormModal (0 = none). Hide IP Limit field entirely when ipLimitEnable is off — disabled fields still take layout space, this collapses Auth(Hysteria) to full width. - inbounds API: new GET /panel/api/inbounds/options that returns just {id, remark, protocol, port, tlsFlowCapable}. Used by the clients page pickers so the dropdown payload stays small on panels with thousands of clients (drops settings JSON, clientStats, streamSettings). Server-side TlsFlowCapable mirrors Inbound.canEnableTlsFlow so the modal no longer needs to parse streamSettings client-side. - clientInfoMsg now shows attached inbound remarks, and getInboundUsages reports the attached client count per inbound. - api-docs: document the new /options endpoint and add tgId / flow to the clients add/update bodies. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(inbounds): keep Node column visible for node-attached inbounds The Node column was bound to hasActiveNode, so disabling every node hid the column even when inbounds were still attached to those nodes — the admin lost the visual cue that those inbounds belonged to a node and would come back when it was re-enabled. Combine hasActiveNode with a new hasNodeAttachedInbound check (any dbInbound with nodeId != null) so the column survives node-disable. * fix(api-docs): accept functional-component icons in EndpointSection AntD-Vue icons (SafetyCertificateOutlined, etc.) are functional components, so the icon prop's type: Object validator was rejecting them with a "Expected Object, got Function" warning at runtime. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * test: cover crypto, random, netsafe, sub helpers, xray equals, websocket hub, node service Adds ~110 unit tests across previously untested packages. Focus on pure-logic and concurrency surfaces where regressions would silently affect users: - util/crypto, util/random: password hashing round-trip, ss2022 key generation, alphabet/length invariants. - util/netsafe: IsBlockedIP edge cases, NormalizeHost validation, SSRF guard with AllowPrivate context bypass. - util/common, util/json_util: traffic formatter, Combine nil-skip, RawMessage empty-as-null and copy-on-unmarshal. - sub: splitLinkLines, searchKey/searchHost, kcp share fields, finalmask normalization, buildVmessLink round-trip. - xray: Config.Equals and InboundConfig.Equals field-by-field, getRequiredUserString/getOptionalUserString type checks. - web/websocket: hub registration, throttling, slow-client eviction, nil-receiver safety, concurrent register/unregister. - web/service: NodeService.normalize validation, normalizeBasePath, HeartbeatPatch.ToUI mapping. - web/job: atomicBool concurrent set/takeAndReset semantics. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * i18n(clients): replace English fallbacks with proper translation keys Pulls every hard-coded English label/title in the Clients page and its four modals through the i18n layer so localized panels stop leaking English. New keys live under pages.clients (auth, hysteriaAuth, uuid, flow, flowNone, reverseTag, reverseTagPlaceholder, telegramId, telegramIdPlaceholder, created, updated, ipLimit) plus refresh at the root and toasts.bulkDeletedMixed / bulkCreatedMixed for partial-failure toasts. Also switches the add-client modal's primary button from "Add" to "Create" for consistency with other create flows. The bulk-add Random/Random+Prefix/... email-method options stay hard-coded by request - they're identifier-shaped strings. * i18n: backfill 99 missing keys across all 12 non-English locales Brings every translation file up to parity with en-US.json so the Clients page, the fallback-children inbound section, the new refresh verb, the Nodes panel-version label and a handful of older holes stop falling through to the English fallback. New strings span: - pages.clients.* (labels, confirmations, toasts, emailMethods) - pages.inbounds.portFallback.* (Reality fallback inbound section) - pages.nodes.panelVersion, menu.clients, refresh Technical identifiers (Auth, UUID, Flow, Reverse tag) are intentionally left untranslated since they correspond to xray-core field names. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * i18n: drop stale pages.client block duplicated in every non-English locale Every non-English locale carried a pages.client (singular) section with 30 entries that duplicated pages.clients (plural). The plural namespace is what the Vue code actually consumes; the singular one was dead weight from an older rename that never got cleaned up in the non-English files. Removing it brings every locale to exactly 984 keys, matching en-US.json. * chore: apply modernize analyzer fixes across codebase Mechanical replacements suggested by golang.org/x/tools/.../modernize: strings.Cut/CutPrefix/SplitSeq, slices.Contains, maps.Copy, min(), range-over-int, new(expr), strings.Builder for hot += loops, reflect.TypeFor[T](), sync.WaitGroup.Go(), drop legacy //+build lines. * feat(database): add PostgreSQL as an optional backend alongside SQLite Lets operators with large client counts or multi-node setups pick PostgreSQL at install time without breaking the existing SQLite default. Backend is selected at runtime via XUI_DB_TYPE/XUI_DB_DSN, a small dialect layer keeps the five JSON_EXTRACT/JSON_EACH queries portable, and a new `x-ui migrate-db` subcommand copies SQLite data into PostgreSQL in FK-aware order. * fix(inbounds): gate node selector to multi-node-capable protocols Hide the Deploy-To selector and clear nodeId when switching to a protocol that can't run on a remote node. Also: - subs: return 404 (not 400) when subId matches no inbounds, so VPN clients distinguish "deleted/unknown" from a server error - hysteria link gen: use the inbound's resolved address so node-managed inbounds advertise the node host instead of the central panel - shadowsocks: default network to 'tcp' (udp was causing issues for some clients on first-create) - vite dev proxy: rewrite migrated-route bypass against the live base path instead of a hardcoded single-segment regex Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): bulk add/delete correctness + perf, working pagination, delayed-start in form Bulk add/delete were serial on the frontend (one toast per call, N round-trips) and the backend race exposed by parallelizing them lost client attachments and hit UNIQUE constraint failed on client_inbounds. The single add/edit modal also had no Start-After-First-Use option, and the table never showed the delayed duration. Backend (web/service/client.go): - Per-inbound mutex on Add/Update/Del InboundClient so concurrent writers on the same inbound don't lose the read-modify-write of settings JSON. - SyncInbound skips create+join when the email is tombstoned so a concurrent maintenance pass (adjustTraffics, autoRenewClients, markClientsDisabledIn- Settings) that did a stale RMW can't resurrect a just-deleted client with a fresh id. - compactOrphans sweeps settings.clients entries whose ClientRecord no longer exists, applied in Add/DelInboundClient + DelInboundClientByEmail so each user-initiated mutation self-heals the inbound's settings. - DelInboundClient uses Pluck instead of First for the stats lookup so a missing row doesn't abort the delete with a noisy ErrRecordNotFound log. Frontend: - HttpUtil.{get,post} accept a silent option that suppresses the auto-toast. - ClientBulkAddModal fires creates in parallel + silent + one summary toast. - useClients.removeMany runs deletes in parallel + silent and refreshes once; ClientsPage bulk delete uses it and shows one aggregate toast. - useClients.applyInvalidate debounces 200 ms so the burst of N WebSocket invalidate events from the backend collapses into a single refresh. - ClientsPage pagination is reactive (paginationState ref + tablePagination computed); onTableChange persists page-size and page changes. - ClientFormModal gains a Start-After-First-Use switch + Duration days input alongside the existing Expiry Date picker; on edit-mode open a negative expiryTime is decoded back to delayed mode + days; on submit the payload sends -86400000 * days or the absolute timestamp. - ClientsPage table shows the delayed-start duration (blue tag Nd, tooltip Start After First Use: Nd) instead of infinity. - Telegram ID field in the form is hidden when /panel/setting/defaultSettings reports tgBotEnable=false; Comment then fills the row. - Form row 3 collapses UUID (span 12) + Total GB (span 8) + Limit IP (span 4) when ipLimitEnable is on, else UUID + Total GB at 12/12. - useInbounds.rollupClients counts only clients with a matching clientStats row, so orphans in settings.clients no longer inflate the inbound's count. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(windows): clean shutdown, working panel restart, harden kernel32 load Three Windows-specific issues addressed: 1. Orphaned xray-windows-amd64 after VS Code debugger stop. Delve's "Stop" sends TerminateProcess to the Go binary, which is uncatchable — our signal handlers never run, so xrayService.StopXray() is skipped and xray is left dangling. Spawn xray as a child of a Job Object with JOB_OBJECT_LIMIT_KILL_ON_JOB_CLOSE so the OS kills xray when our handle to the job is closed (which happens even on TerminateProcess). Also trap os.Interrupt in main so Ctrl+C in the terminal runs the graceful path. 2. /panel/setting/restartPanel logged "failed to send SIGHUP signal: not supported by windows" because Windows can't deliver arbitrary signals. Add a restart hook in web/global; main registers it to push SIGHUP into its own signal channel, and RestartPanel calls the hook before falling back to the (Unix-only) signal path. Same restart-loop code runs in both cases. 3. util/sys/sys_windows.go now uses windows.NewLazySystemDLL so the kernel32.dll resolve is pinned to %SystemRoot%\System32 (prevents DLL hijacking by a planted DLL next to the binary). Local filetime type replaced with windows.Filetime, and the unreliable syscall.GetLastError() fallback replaced with a type assertion on the errno captured at call time. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(sys): correct CPU/connection accounting on linux + darwin util/sys/sys_linux.go: - GetTCPCount/GetUDPCount were counting the column header row in /proc/net/{tcp,udp}[6] as a connection, inflating the reported total by 1 per non-empty file (so the panel status line always showed 2 more connections than actually existed). Replace getLinesNum + safeGetLinesNum with a single bufio.Scanner-based countConnections that skips the header. - CPUPercentRaw now opens HostProc("stat") instead of a hardcoded /proc/stat so HOST_PROC overrides apply, matching the connection counters in the same file. - Simplify CPU field unpacking: pad nums to 8 once instead of guarding every assignment with a len check. util/sys/sys_darwin.go: - Fix swapped idle/intr indices on kern.cp_time. BSD CPUSTATES order is user, nice, sys, intr, idle (CP_INTR=3, CP_IDLE=4) — gopsutil's cpu_darwin_nocgo.go reads the same layout. The previous code used out[3] as idle and out[4] as intr, so busy = total - dIdle was actually subtracting interrupt time, making the panel report CPU usage close to 100% on macOS regardless of actual load. - Collapse the per-field delta math into a single loop. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(xray): rotate crash reports into log folder, prevent overwrites writeCrashReport had two flaws: it wrote to the bin folder (alongside the xray binary) which conflates artifacts, and the second-precision timestamp meant a tight restart-loop crash burst overwrote prior reports. Write to the log folder with nanosecond precision and keep the last 10 reports. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * revert(inbounds): drop unreleased portfallback protocol The Port-with-Fallback inbound (commit 62fd9f9d) was confusing as a standalone protocol — fallbacks belong on a regular VLESS/Trojan TCP-TLS inbound, the way Xray models them natively. Rip out the entire feature cleanly (no migration needed since it was never released): protocol constant, fallback children DB table, FallbackService, 2 API endpoints, all UI rows, related translations and api-docs. A native fallback flow attached to VLESS/Trojan TCP-TLS/Reality will land in a follow-up commit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds): native fallbacks on VLESS/Trojan TCP-TLS, with working child links A VLESS or Trojan inbound on TCP with TLS or Reality can now act as a fallback master: pick existing inbounds as children and the panel auto- fills the SNI / ALPN / path / xver routing fields from each child's transport, auto-builds settings.fallbacks at config-gen time, and rewrites the child's client-share link so it advertises the master's reachable endpoint and TLS state instead of the child's loopback listen. Layout matches the Xray All-in-One Nginx example: master at :443 with clients + TLS, each child on 127.0.0.1 with its own transport+clients. Order matters (Xray walks fallbacks top-to-bottom) — reorder via the per-row up/down arrows. Path / SNI / ALPN are exposed under a per-row Edit toggle for the rare cases where the auto-derivation needs overriding; otherwise just pick a child and you're done. Backend: new InboundFallback table + FallbackService (GetByMaster / SetByMaster / GetParentForChild / BuildFallbacksJSON); two routes (GET / POST /panel/api/inbounds/:id/fallbacks); xray.GetXrayConfig injects settings.fallbacks for any VLESS/Trojan TCP-TLS/Reality inbound; GetInbounds annotates each child with FallbackParent so the frontend can rewrite links without an extra round-trip. Link projection covers every emission path — clients-page QR/links, per-inbound Get URL, raw subscription, sub-JSON, sub-Clash, and the inbounds-page link/info/QR — via a shared projectThroughFallbackMaster on the backend and a shared projectChildThroughMaster on the frontend that both handle the panel-tracked relationship and the legacy unix-socket (@vless-ws) convention. Strings translated into all 12 non-English locales. * docs: rewrite CONTRIBUTING with full local-dev setup The prior three-line CONTRIBUTING left newcomers guessing at every non-trivial step: which Go / Node versions, where xray comes from, why the panel goes blank when XUI_DEBUG=true is flipped on, how the Vue multi-page setup is wired, what to do on Windows when go build trips on the CGo SQLite driver. Now covers prerequisites, MinGW-w64 install on Windows (niXman builds or MSYS2), one-shot first-time setup, two frontend dev workflows with the XUI_DEBUG asset-cache gotcha called out, the architecture and conventions of the Vue side, a project-layout map, useful env vars, and the PR checklist. ---------
2026-05-19 03:16:42 -07:00
local existing_webBasePath=$(${xui_folder}/x-ui setting -show true | grep -Eo 'webBasePath: .+' | awk '{print $2}' | sed 's#^/##')
2026-05-19 04:09:35 -07:00
local existing_port=$(${xui_folder}/x-ui setting -show true | grep -Eo 'port: .+' | awk '{print $2}')
# Properly detect empty cert by checking if cert: line exists and has content after it
local existing_cert=$(${xui_folder}/x-ui setting -getCert true | grep 'cert:' | awk -F': ' '{print $2}' | tr -d '[:space:]')
2025-08-21 05:24:25 -07:00
local URL_lists=(
"https://api4.ipify.org"
"https://ipv4.icanhazip.com"
"https://v4.api.ipinfo.io/ip"
"https://ipv4.myexternalip.com/raw"
"https://4.ident.me"
"https://check-host.net/ip"
2025-08-21 05:24:25 -07:00
)
local server_ip=""
for ip_address in "${URL_lists[@]}"; do
2026-05-04 04:20:24 -07:00
local response=$(curl -s -w "\n%{http_code}" --max-time 3 "${ip_address}" 2> /dev/null)
local http_code=$(echo "$response" | tail -n1)
local ip_result=$(echo "$response" | head -n-1 | tr -d '[:space:]"')
if [[ "${http_code}" == "200" && "${ip_result}" =~ ^[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
server_ip="${ip_result}"
break
fi
done
2026-05-04 04:20:24 -07:00
if [[ -z "$server_ip" ]]; then
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ "$NONINTERACTIVE" == "1" ]]; then
# Panel binds 0.0.0.0 regardless; the IP is only used to compose the
# displayed access URL. Fall back to XUI_SERVER_IP or leave blank.
server_ip="${XUI_SERVER_IP:-}"
else
echo -e "${yellow}Could not auto-detect server IP from any provider.${plain}"
while [[ -z "$server_ip" ]]; do
read -rp "Please enter your server's public IPv4 address: " server_ip
server_ip="${server_ip// /}"
if [[ ! "$server_ip" =~ ^[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
echo -e "${red}Invalid IPv4 address. Please try again.${plain}"
server_ip=""
fi
done
fi
fi
2024-10-25 02:05:22 -07:00
if [[ ${#existing_webBasePath} -lt 4 ]]; then
2026-05-19 04:09:35 -07:00
if [[ "$existing_hasDefaultCredential" == "true" ]]; then
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
local config_webBasePath="${XUI_WEB_BASE_PATH:-$(gen_random_string 18)}"
local config_username="${XUI_USERNAME:-$(gen_random_string 10)}"
local config_password="${XUI_PASSWORD:-$(gen_random_string 10)}"
local config_port=""
2026-05-04 04:20:24 -07:00
local db_label="SQLite (/etc/x-ui/x-ui.db)"
echo ""
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${green} Database Selection ${plain}"
echo -e "${green}═══════════════════════════════════════════${plain}"
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
echo -e " 1) SQLite (default — recommended for < 500 clients)"
echo -e " 2) PostgreSQL (recommended for high client counts / many nodes)"
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ "$NONINTERACTIVE" == "1" ]]; then
if [[ "${XUI_DB_TYPE:-sqlite}" == "postgres" ]]; then
db_choice="2"
else
db_choice="1"
fi
else
read -rp "Choose [1]: " db_choice
db_choice="${db_choice:-1}"
fi
if [[ "$db_choice" == "2" ]]; then
local xui_env_file
case "${release}" in
ubuntu | debian | armbian)
xui_env_file="/etc/default/x-ui"
;;
arch | manjaro | parch | alpine)
xui_env_file="/etc/conf.d/x-ui"
;;
*)
xui_env_file="/etc/sysconfig/x-ui"
;;
esac
local xui_dsn=""
local pg_mode=""
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
local pg_local_installed=0
while [[ -z "$xui_dsn" ]]; do
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ "$NONINTERACTIVE" == "1" ]]; then
if [[ -n "${XUI_DB_DSN:-}" ]]; then
xui_dsn="${XUI_DB_DSN}"
db_label="PostgreSQL (external)"
break
fi
echo -e "${yellow}Installing PostgreSQL locally (non-interactive)...${plain}"
local pg_cred_file
pg_cred_file=$(mktemp 2> /dev/null) || pg_cred_file=$(mktemp -t x-ui-pg-creds.XXXXXXXX)
if [[ -n "${pg_cred_file}" ]] && xui_dsn=$(PG_CRED_FILE="${pg_cred_file}" install_postgres_local); then
pg_local_installed=1
if [[ -r "${pg_cred_file}" ]]; then
# shellcheck disable=SC1090
source "${pg_cred_file}"
fi
rm -f "${pg_cred_file}"
db_label="PostgreSQL (${PG_USER}@${PG_HOST}:${PG_PORT}/${PG_DB})"
break
fi
rm -f "${pg_cred_file}"
echo -e "${red}PostgreSQL installation failed in non-interactive mode; aborting.${plain}"
echo -e "${yellow}Set XUI_DB_DSN to use an existing server, or XUI_DB_TYPE=sqlite.${plain}"
exit 1
fi
echo ""
echo -e " 1) Install PostgreSQL locally and create a dedicated user/db (recommended)"
echo -e " 2) Use an existing PostgreSQL server (enter DSN)"
read -rp "Choose [1]: " pg_mode
pg_mode="${pg_mode:-1}"
if [[ "$pg_mode" == "2" ]]; then
while [[ -z "$xui_dsn" ]]; do
read -rp "Enter PostgreSQL DSN (postgres://user:pass@host:port/dbname?sslmode=disable): " xui_dsn
xui_dsn="${xui_dsn// /}"
done
db_label="PostgreSQL (external)"
else
echo -e "${yellow}Installing PostgreSQL — this may take a moment...${plain}"
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
local pg_cred_file
pg_cred_file=$(mktemp 2> /dev/null) || pg_cred_file=$(mktemp -t x-ui-pg-creds.XXXXXXXX)
if [[ -z "${pg_cred_file}" ]]; then
echo -e "${red}Failed to create temporary credentials file.${plain}"
xui_dsn=""
continue
fi
if xui_dsn=$(PG_CRED_FILE="${pg_cred_file}" install_postgres_local); then
pg_local_installed=1
if [[ -r "${pg_cred_file}" ]]; then
# shellcheck disable=SC1090
source "${pg_cred_file}"
fi
rm -f "${pg_cred_file}"
db_label="PostgreSQL (${PG_USER}@${PG_HOST}:${PG_PORT}/${PG_DB})"
else
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
rm -f "${pg_cred_file}"
echo ""
echo -e "${red}PostgreSQL installation failed.${plain}"
echo -e " 1) Retry local install"
echo -e " 2) Enter an external DSN instead"
echo -e " 3) Abort install"
echo -e " 4) Fall back to SQLite"
read -rp "Choose [1]: " pg_fail
pg_fail="${pg_fail:-1}"
case "$pg_fail" in
2) pg_mode="2" ;;
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
3)
echo -e "${red}Install aborted.${plain}"
exit 1
;;
4)
db_choice="1"
xui_dsn=""
break
;;
*) xui_dsn="" ;;
esac
fi
fi
done
if [[ -n "$xui_dsn" ]]; then
install -d -m 755 "$(dirname "$xui_env_file")"
umask 077
cat > "$xui_env_file" << EOF
XUI_DB_TYPE=postgres
XUI_DB_DSN=${xui_dsn}
EOF
chmod 600 "$xui_env_file"
umask 022
export XUI_DB_TYPE=postgres
export XUI_DB_DSN="${xui_dsn}"
ensure_pg_client || echo -e "${yellow}⚠ Could not install pg_dump/pg_restore. In-panel database backup/restore will be unavailable until you install the postgresql-client package.${plain}"
fi
fi
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
if [[ "$NONINTERACTIVE" == "1" ]]; then
if [[ -n "${XUI_PANEL_PORT:-}" ]]; then
config_port="${XUI_PANEL_PORT}"
echo -e "${yellow}Your Panel Port is: ${config_port}${plain}"
else
config_port=$(shuf -i 1024-62000 -n 1)
echo -e "${yellow}Generated random port: ${config_port}${plain}"
fi
Feat/multi inbound clients (#4469) * feat(clients): add shadow tables for first-class client promotion Introduces three new GORM-backed tables (clients, client_inbounds, inbound_fallback_children) and a populate-only seeder that backfills them from each inbound's existing settings.clients JSON. Duplicate emails across inbounds auto-merge under one client row, with each field conflict logged. Existing services are unchanged and continue reading from settings.clients — this commit is groundwork only. * feat(clients): make clients+client_inbounds the runtime source of truth Adds ClientService.SyncInbound that reconciles the new tables from each inbound's clients list whenever existing service paths mutate settings.clients. Wires it into AddInbound, UpdateInbound, AddInboundClient, UpdateInboundClient, DelInboundClient, DelInboundClientByEmail, DelDepletedClients, autoRenewClients, and the timestamp-backfill path in adjustTraffics, plus DetachInbound on DelInbound. GetXrayConfig now builds settings.clients from the new tables before writing config.json, and getInboundsBySubId joins through them instead of JSON_EACH on settings JSON. Live Xray config and subscription endpoints are now driven by the relational view; settings.clients JSON stays in step as a side effect of every write. * feat(clients): add top-level Clients tab and CRUD API Adds /panel/api/clients endpoints (list, get, add, update, del, attach, detach) backed by ClientService methods that orchestrate the per-inbound Add/Update/Del flows so a single client row is created once and attached to many inbounds in one operation. The frontend gains a dedicated Clients page (frontend/clients.html + src/pages/clients/) with an AntD table, multi-inbound attach modal, and full CRUD. Axios interceptor learns to honour Content-Type: application/json so the JSON endpoints work alongside the legacy form-encoded ones. The legacy per-inbound client modal stays untouched in this PR — both flows now write to the same source of truth. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds): add Port-with-Fallback inbound type Adds a new "portfallback" protocol that emits as a VLESS-TLS inbound under the hood but is paired with a sidecar table of child inbounds. Panel auto-builds settings.fallbacks at Xray-config-gen time from the sidecar — each child's listen+port becomes the fallback dest, with SNI/ALPN/path/xver match criteria pulled from the row. No more typing loopback ports by hand or keeping settings.fallbacks in sync. Backend: new FallbackService (Get/SetChildren, BuildFallbacksJSON); two new routes (GET/POST /panel/api/inbounds/:id/fallbackChildren); xray.GetXrayConfig injects fallbacks for PortFallback inbounds; the inbound model emits protocol="vless" so Xray accepts the config. Frontend: PORTFALLBACK joins the protocol dropdown; selecting it shows the standard VLESS controls plus a Fallback Children table (inbound picker + per-row SNI/ALPN/path/xver). Children are loaded on edit and replaced atomically on save. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Reset Traffic, QR Code, Info actions + Online/Remaining columns The Clients page table gains: - Online column — green/grey tag driven by /panel/api/inbounds/onlines, polled every 10s. - Remaining column — bytes-remaining tag, coloured green/orange/red against quota, purple infinity when unlimited. - Action icons per row: QR, Info, Reset traffic, Edit, Delete. ClientInfoModal shows the full client detail (uuid/password/auth, traffic ↑/↓ + remaining + all-time, expiry absolute + relative, attached inbounds chip list, online + last-online). ClientQrModal fetches links for the client's subId via /panel/api/inbounds/getSubLinks/:subId and renders each one through the existing QrPanel component. Reset Traffic confirms then calls the existing per-inbound endpoint on the client's first attached inbound (the traffic row is keyed on email globally, so any attached inbound resets the shared counter). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): expose Attached inbounds in edit mode The multi-select was gated on add-only, so editing a client had no way to change which inbounds it belonged to. The picker now shows in both modes, and on submit the modal diffs the picked set against the original attachedIds — additions go through the /attach endpoint, removals through /detach, both after the field update lands so the new attachments get the latest values. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): unbreak template parsing + stale i18n keys - InboundFormModal: split the multi-line help string in the PortFallback section onto one line — Vue's template parser was bailing on Unterminated string constant because a single-quoted literal spanned two lines inside a {{ }} interpolation. - ClientInfoModal: t('disable') was missing at the root level, so vue-i18n returned the key path literally. Use t('disabled') which exists. - Linter cleanup elsewhere: pages.client.* references renamed to pages.clients.* to match the merged i18n block; whitespace normalisation in a few unrelated Vue templates. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * 1 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(traffic): drop all-time traffic tracking Removes the AllTime field from Inbound and ClientTraffic and migrates existing DBs by dropping the all_time columns on startup. The counter duplicated up+down without adding signal, and the per-event accumulator ran on every traffic write. Frontend: drop the All-time column from the inbound list and the client-row table, the All-time row from the client info modal, and the All-Time Total Usage tile from the inbounds summary card. The allTimeTraffic/allTimeTrafficUsage i18n keys are removed across every locale. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): mobile cards, multi-select, bulk add Adds the same row-card layout the inbounds page uses on mobile: the table is suppressed under the mobile breakpoint and each client renders as a compact card with a status dot, email, Info button, Enable switch, and overflow menu. All the per-client detail (traffic, remaining, expiry, attached inbounds, flow, created/updated, URL, subscription) opens through the existing info modal. Multi-select with bulk delete wires AntD row-selection on desktop and a per-card checkbox on mobile; a Delete (N) button appears in the toolbar when anything is selected. Bulk add reuses the five email-generation modes from the inbound bulk modal but takes a multi-inbound picker so one bulk run can attach to several inbounds at once. Submits client-by-client through the existing /panel/api/clients/add endpoint. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(inbounds): remove legacy per-inbound client UI Now that clients live as first-class rows attached to one or many inbounds, the per-inbound client UI on the inbounds page is dead weight — every client action either has a global equivalent on the Clients page or makes no sense in a many-to-many world. Deletes ClientFormModal, ClientBulkModal, CopyClientsModal, and ClientRowTable from inbounds/. Strips the matching emits, refs, handlers, and dropdown menu items from InboundList and InboundsPage, and removes the dead mobile expand-chevron state and the desktop expanded-row plumbing that drove the inline client table. The InboundFormModal Clients tab still works in add-mode (one inline client at inbound creation) — that flow goes through ClientService. SyncInbound on save and remains useful. Fixes a stray "</a-dropdown>" left over by an earlier toolbar edit in ClientsPage that broke the template parser. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Delete depleted action Mirrors the legacy delDepletedClients action that lived under the inbounds page, but as a first-class /panel/api/clients/delDepleted endpoint backed by ClientService. The new path goes through ClientService.Delete for each depleted email, so the new clients + client_inbounds + xray_client_traffic tables stay consistent. Adds a danger-styled toolbar button on the Clients page (next to Reset all client traffic) with a confirm dialog and a toast reporting the deleted count. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(api): move every client-shaped endpoint off /inbounds onto /clients After the multi-inbound client migration, client state belongs to the client API surface, not the inbound one. Twelve routes that were crammed under /panel/api/inbounds/* now live where they belong, under /panel/api/clients/*. Moved (route, handler, doc): POST /clientIps/:email POST /clearClientIps/:email POST /onlines POST /lastOnline POST /updateClientTraffic/:email POST /resetAllClientTraffics/:id POST /delDepletedClients/:id POST /:id/resetClientTraffic/:email GET /getClientTraffics/:email GET /getClientTrafficsById/:id GET /getSubLinks/:subId GET /getClientLinks/:id/:email Their /clients/* counterparts are: POST /clients/clientIps/:email POST /clients/clearClientIps/:email POST /clients/onlines POST /clients/lastOnline POST /clients/updateTraffic/:email POST /clients/resetTraffic/:email (email-only, fans out) GET /clients/traffic/:email GET /clients/traffic/byId/:id GET /clients/subLinks/:subId GET /clients/links/:id/:email per-inbound resetAllClientTraffics and delDepletedClients are dropped entirely — the Clients page already exposes global Reset All Traffic and Delete depleted actions, and per-inbound resets are meaningless once a client can be attached to many inbounds. ClientService.ResetTrafficByEmail is the new email-only reset path: it looks up every inbound the client is attached to and pushes the counter reset + Xray re-add through inboundService.ResetClientTraffic for each one, so depleted users come back online instantly. Frontend callers (ClientsPage, useClients, ClientQrModal, ClientInfoModal, InboundInfoModal, InboundsPage, useInbounds) all switched to the new paths. The Inbounds page drops its per-inbound "Reset client traffic" and "Delete depleted clients" dropdown items — users do those at the client level now. api-docs is rebuilt to match. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(service): switch tgbot + ldap callers to ClientService Adds two thin helpers to ClientService (CreateOne, DetachByEmail) and rewrites tgbot.SubmitAddClient and ldap_sync_job to call ClientService directly. Removes the JSON-blob payloads (BuildJSONForProtocol output for add, clientsToJSON/clientToJSON helpers) that callers previously fed to InboundService.AddInboundClient/DelInboundClient. ldap_sync_job.batchSetEnable now loops InboundService.SetClientEnableByEmail per email instead of trying to coerce AddInboundClient into doing the update — the old path would have failed duplicate-email validation for existing clients anyway. The legacy InboundService.AddInboundClient/UpdateInboundClient/ DelInboundClient methods stay in place; they are now only used internally by ClientService Create/Update/Delete/Attach. Inlining + deleting them follows in a separate commit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(service): move all client mutation methods to ClientService Moves the client mutation surface out of InboundService and into ClientService. These methods all operate on a single client (identity fields, traffic limits, expiry, ip limit, enable state, telegram tg id) and didn't belong on the inbound aggregate. Moved (12 methods): AddInboundClient, UpdateInboundClient, DelInboundClient, DelInboundClientByEmail, checkEmailsExistForClients, SetClientTelegramUserID, checkIsEnabledByEmail, ToggleClientEnableByEmail, SetClientEnableByEmail, ResetClientIpLimitByEmail, ResetClientExpiryTimeByEmail, ResetClientTrafficLimitByEmail. Each method now takes an explicit *InboundService for the helpers that legitimately stay on InboundService (GetInbound, GetClients, runtimeFor, AddClientStat / UpdateClientStat / DelClientStat, DelClientIPs / UpdateClientIPs, emailUsedByOtherInbounds, getAllEmailSubIDs, GetClientInboundByEmail / GetClientInboundByTrafficID, GetClientTrafficByEmail). Stays on InboundService: ResetClientTrafficByEmail and ResetClientTraffic(id, email) — these mutate xray_client_traffic rows, not client identity, so they're inbound-side bookkeeping. Callers updated: tgbot (6 calls), ldap_sync_job (1 call), InboundService internal (writeBackClientSubID, CopyInboundClients, AddInbound's email-uniqueness check), ClientService Create/Update/ Delete/Attach/Detach. Also removes a dead resetAllClientTraffics controller handler whose route was already gone after the previous /clients API migration. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(clients): finish migrating to ClientService + tidy IP routes Two related cleanups in the new /clients surface: 1. Move ResetAllClientTraffics (bulk-reset of xray_client_traffic + last_traffic_reset_time, with node-runtime propagation) from InboundService to ClientService. PeriodicTrafficResetJob now holds a clientService and calls j.clientService.ResetAllClientTraffics(&j.inboundService, id). The last client-mutation method on InboundService is gone. 2. Shorten redundantly-named routes/handlers under /panel/api/clients: - /clientIps/:email -> /ips/:email (handler getIps) - /clearClientIps/:email -> /clearIps/:email (handler clearIps) The "client" prefix was redundant inside the clients namespace. Frontend (InboundInfoModal) and api-docs updated to match. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds,clients): clean up inbound modal + enrich client modal Inbound modal rework (InboundFormModal.vue + inbound.js): - Drop the embedded Client subform in the Protocol tab. Multi-inbound clients are managed exclusively from the Clients page now; a fresh inbound is created with zero clients (settings constructors default to []) and the user attaches clients afterwards. - Hide the Protocol tab entirely when it has nothing to render (VMESS, Trojan without fallbacks, Hysteria). Auto-switches active tab to Basic when the tab disappears while focused. - Move the Security section (Security selector + TLS block with certs and ECH + Reality block) out of the Stream tab into its own Security tab, sharing the canEnableStream gate. Client modal additions (ClientFormModal.vue + ClientBulkAddModal.vue): - Flow select (xtls-rprx-vision / -udp443) appears only when the panel actually has a Vision-capable inbound (VLESS or PortFallback on TCP with TLS or Reality). Hidden otherwise, and cleared when it disappears. - IP Limit input is disabled when the panel-level ipLimitEnable setting is off, fetched into useClients alongside subSettings and threaded through ClientsPage to both modals. - Edit modal now shows an "IP Log" section listing IPs that have connected with the client's credentials, with refresh and clear buttons (calls the renamed /panel/api/clients/ips and /clearIps endpoints). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(inbounds): drop manual Fallbacks UI from inbound modal The PortFallback protocol type now covers the common VLESS-master-plus-children case with auto-wired dests, so the manual Fallbacks editor (showFallbacks block in the Protocol tab) is mostly redundant. Removed: - the v-if="showFallbacks" template block (SNI/ALPN/Path/dest/PROXY rows) - the showFallbacks computed - the addFallback / delFallback helpers - the .fallbacks-header / .fallbacks-title styles - the showFallbacks gate from hasProtocolTabContent (so Trojan-over-TCP no longer shows an empty Protocol tab) Power users who need a non-inbound fallback dest (nginx, static site) can still author settings.fallbacks via the Advanced JSON tab. * feat(clients,inbounds): move search/filter to Clients page + small fixes Search/filter relocation: - Remove the search/filter toolbar (search switch + filter radio + protocol/node selects + the visibleInbounds projection + inboundsFilterState localStorage + filter CSS + the SearchOutlined/ FilterOutlined/ObjectUtil/Inbound imports it required) from InboundList. The filters were all client-oriented buckets bolted onto the inbound row. - Add a search/filter toolbar to ClientsPage with the same shape: switch between deep-text search and bucket filter (active / deactive / depleted / expiring / online) + protocol filter that matches clients attached to at least one inbound with the chosen protocol. State persists in clientsFilterState localStorage. filteredClients drives both the desktop table and the mobile card list, and select-all / allSelected / someSelected only span the visible subset. - useClients now also fetches expireDiff and trafficDiff from /panel/setting/defaultSettings (used to detect the expiring bucket); ClientsPage threads them into the client-bucket helper. Loose fixes folded in: - Add Client: email field is auto-filled with a random handle on open, matching uuid/subId/password/auth. - Inbound clone: parse and reuse the source settings JSON (with clients reset to []) instead of building a fresh defaulted Settings, so VLESS Encryption/Decryption and other non-client fields survive the clone. - en-US.json: add the ipLog string used by the edit-client modal. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Reverse tag field for VLESS-attached clients Mirrors the Flow field's pattern: a Reverse tag input appears in the Add/Edit Client modal whenever at least one selected inbound is VLESS or PortFallback. The value rides over the wire as client.reverse = { tag: '...' } so it lands directly in model.Client's *ClientReverse field; an empty value omits the reverse key entirely. On edit the field is hydrated from props.client.reverse?.tag, and the showReverseTag watcher clears the field if the user drops the last VLESS-like inbound from the selection. * fix(xray): emit only protocol-relevant fields per client entry The Xray config synthesizer was writing every identifier field (id, password, flow, auth, security/method, reverse) on every client entry regardless of the inbound's protocol. Xray ignores unknown fields, so the config worked, but it diverged from the spec and leaked secrets across protocols when one client was attached to multiple inbounds — a VLESS inbound's generated config carried the same client's Trojan password and Hysteria auth alongside its uuid. Switch on inbound.Protocol when building each entry: - VLESS / PortFallback: id, flow, reverse - VMess: id, security - Trojan: password, flow - Shadowsocks: password, method - Hysteria / Hysteria2: auth email is emitted for every protocol. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): restore auto-disable kick under new schema disableInvalidClients still resolved (inbound_tag, email) pairs via JSON_EACH(inbounds.settings.clients), which is empty after migrating to the clients + client_inbounds tables. Result: xrayApi.RemoveUser never ran for depleted clients, clients.enable stayed true so the UI showed them as active, and only xray_client_traffic.enable got flipped - making "Restart Xray After Auto Disable" only half-work. Resolve the targets via a JOIN through the new schema, flip clients.enable so the Clients page reflects the state, and drop the legacy JSON write-back plus the subId cascade workaround (email is unique now). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): live WebSocket updates + Ended status surfacing ClientsPage now subscribes to traffic / client_stats / invalidate WebSocket events instead of polling /onlines every 10s. Per-row traffic counters refresh in place, online state stays current, and list-level mutations elsewhere trigger a refresh. The client roll-up summary moves from InboundsPage to ClientsPage where it belongs, restructured into six labeled stat tiles (Total / Online / Ended / Expiring / Disabled / Active) with email popovers on the ones with issues. Auto-disabled clients (traffic exhausted or expiry passed) now classify as 'depleted' even though clients.enable=false, so they show up under the Ended filter and render a red Ended tag instead of looking indistinguishable from an operator-disabled row. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(nodes): per-node client roll-up and panel version Added transient inboundCount / clientCount / onlineCount / depletedCount fields to model.Node, populated by NodeService.GetAll via aggregated queries (one join across inbounds + client_inbounds, one over client_traffics intersected with the in-memory online emails). The Nodes list renders these as colored chips on a new "Clients" column so an operator can see at a glance how many users each node carries and how many are currently online or depleted. Also exposes the remote panel's version. The central panel adds panelVersion to its /api/server/status payload (sourced from config.GetVersion). Probe reads that field and persists it on the node row, mirroring how xrayVersion already flows. NodesPage gets a new column next to Xray Version, in both desktop and mobile views, with English and Persian strings. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): stop node sync from resurrecting deleted clients Several related issues around node-managed clients: - Remote runtime: drop the per-inbound resetAllClientTraffics path and point traffic/onlines/lastOnline fetches at the new /panel/api/clients/* routes. - Delete from master: always push the updated inbound to the node even when the client was already disabled or depleted, so the node actually loses the user instead of silently keeping it. - setRemoteTraffic: mirror remote clients into the central tables only on first discovery of a node inbound. Matched inbounds let the master own the join table, so a stale snap can no longer re-create a ClientRecord (and join row) for a client that was just deleted on the master. - ClientService.Delete: route through submitTrafficWrite so deletes serialize with node traffic merges, and switch the final ClientRecord delete to an explicit Where("id = ?") clause. - setRemoteTraffic UNIQUE-constraint fix: use clause.OnConflict on inserts and email-keyed UPDATEs for client_traffics, so mirroring a snap doesn't trip the unique email index. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(clients): switch client API endpoints from id to email All client-scoped routes now use the unique email as the path key (get, update, del, attach, detach, links). Email is the stable, protocol-independent identifier — UUIDs don't exist for trojan or shadowsocks, and internal numeric ids leaked panel implementation detail into the public API. Removed the redundant /traffic/byId/:id endpoint (covered by /traffic/:email) and collapsed /links/:id/:email into /links/:email, which now returns links across every attached inbound for the client. Frontend selection, bulk delete, and toggle state are now keyed by email as well, dropping the id→email lookup workaround. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(server): move cached state and helpers into ServerService ServerController had grown to hold its own status cache, version-list TTL cache, history-bucket whitelist, and the loop that drove all three — concerns that belong in the service layer. Pull them out: - lastStatus + the @2s refresh become ServerService.RefreshStatus and ServerService.LastStatus; the controller's cron now just orchestrates the cross-service side effects (xrayMetrics sample, websocket broadcast). - The 15-minute Xray-versions cache (with stale-on-error fallback) moves into ServerService.GetXrayVersionsCached, collapsing the controller handler to a single call. - The freedom/blackhole outbound-tag walk used by /xraylogs becomes ServerService.GetDefaultLogOutboundTags. - The allowed-history-bucket whitelist moves to package-level service.IsAllowedHistoryBucket, so both NodeController and ServerController validate against the same list. Net result: web/controller/server.go drops from 458 to 365 lines and contains only HTTP wiring + presentation-y side effects. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(api): emit JSON-text columns as nested objects Inbound, ClientRecord, and InboundClientIps store settings / streamSettings / sniffing / reverse / ips as JSON-text in the DB. The API was passing that text through verbatim, so every consumer had to JSON.parse a string inside a string. Add MarshalJSON / UnmarshalJSON so the wire format is a real nested object, while still accepting the legacy escaped-string shape on write. Frontend dbinbound.js gets a matching coerceInboundJsonField helper for the same dual-shape read path, and inbound.js toJson stops emitting empty/placeholder fields (externalProxy [], sniffing destOverride when disabled, etc.) so the new normalised JSON stays terse. api-docs and the inbound-clone path are updated to the new shape. Controller route lists are regrouped so all GETs sit above POSTs. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): include inboundIds and traffic in /clients/list ClientRecord got its own MarshalJSON in the previous commit, and ClientWithAttachments embeds it to add inboundIds and traffic. Go promotes the embedded MarshalJSON to the outer struct, so the encoder was calling ClientRecord.MarshalJSON for the whole value and silently dropping the extras. The frontend reads row.inboundIds / row.traffic from /clients/list, so attached inbounds didn't render and newly added clients looked like they hadn't saved. Add an explicit MarshalJSON on ClientWithAttachments that splices the extras in. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): gate IP Log on ipLimitEnable + clean access-log dropdown Legacy panel hid the IP Log section when access logging was off; the Vue 3 migration left it gated on isEdit only, so the section showed even when xray's access log was 'none' and nothing was being recorded. Restore the ipLimitEnable gate on the edit modal's IP Log form-item. While here, clean up the Xray Settings access-log dropdown: previously two 'none' entries appeared (an empty value labelled with t('none') and the literal 'none' from the options array). Drop the empty option for access log (the literal 'none' covers it) and relabel the empty option for error log / mask address to t('empty') so they're distinguishable. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(nodes): route per-client ops through node clients API + orphan sweep Adds Runtime methods AddClient, UpdateUser, and DeleteUser so master mutates clients on a node via /panel/api/clients/{add,update,del} rather than pushing the whole inbound. The previous rt.UpdateInbound path made the node DelInbound+AddInbound on every single-client change, briefly cycling every other user on the same inbound. DelInbound no longer filters by enable=true, so a disabled node inbound actually gets removed from the node instead of being resurrected by the next snap. setRemoteTrafficLocked now sweeps any ClientRecord with zero ClientInbound rows after SyncInbound rebuilds the attachments, which is how a node-side delete propagates back to master instead of leaving a detached ghost. ClientService.Delete tombstones the email first so a snap arriving mid-delete can't re-create the record. WebSocket broadcasts an "invalidate(clients)" message on every client mutation so the Clients page refreshes without manual reload. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(balancers): allow fallback on all strategies + feed burstObservatory from random/roundRobin Drops the random/roundRobin gate on the Fallback field in BalancerFormModal so every strategy can pick a fallback outbound. syncObservatories now feeds burstObservatory from leastLoad + random + roundRobin balancers (was leastLoad only), matching how leastPing feeds observatory. Fix the JsonEditor "Unexpected end of JSON input" that appeared when switching a balancer between leastPing and another strategy: the obsView watcher was gated on showObsEditor (a boolean OR of the two flags) and missed the case where one observatory swapped for the other in the same tick. Watch the individual flags instead so obsView flips to the surviving editor and the getter stops pointing at a deleted key. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(inbounds): use sortedInbounds for mobile empty-state check InboundList referenced an undefined visibleInbounds in the mobile card list's empty-state guard, throwing "Cannot read properties of undefined (reading 'length')" and breaking the entire mobile render. * feat(clients): sortable table columns Adds the same sortState / sortableCol / sortFns pattern InboundList uses, wrapping filteredClients in sortedClients so sort composes with the existing search/filter pipeline. Sortable: enable, email, inboundIds (attachment count), traffic, remaining, expiryTime; actions and online stay unsorted. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(shadowsocks): generate valid ss2022 keys and per-client method for legacy ciphers The Add Client flow on shadowsocks inbounds was producing xray configs that failed to start: - 2022-blake3-* ciphers need a base64-encoded key of an exact byte length per cipher. fillProtocolDefaults was assigning a uuid-style string, which xray rejects as "bad key". Now the password is generated (or replaced if invalid) via random.Base64Bytes(n) sized to the chosen cipher. - Legacy ciphers (aes-256-gcm, chacha20-*, xchacha20-*) require a per-client method field in multi-user mode; model.Client has no Method, so settings.clients was stored without one and xray failed with "unsupported cipher method:". applyShadowsocksClientMethod now injects the top-level method into each client on add/update, and healShadowsocksClientMethods backfills it at xray-config-build time so existing inbounds heal on the next start. - xray/api.go ssCipherType switch was missing aes-256-gcm, which fell through to ss2022 path. - SSMethods dropdown now offers aes-256-gcm. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): preserve ClientRecord on inbound delete + filter Attached inbounds to multi-client protocols Replace the global orphan sweep in setRemoteTrafficLocked with a per-inbound diff cleanup: only delete a ClientRecord whose email disappeared from a snap-tracked inbound (i.e. a node-side delete). Inbounds that vanished entirely from the snap (e.g. admin deleted the inbound on master) aren't iterated, so a client whose last attachment came from that inbound is now left alone instead of being deleted alongside the inbound. ClientFormModal and ClientBulkAddModal now filter the Attached inbounds dropdown to protocols that actually support multiple clients: shadowsocks, vless, vmess, trojan, hysteria, hysteria2, and portfallback (which routes through VLESS settings). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): make empty-state text readable on dark/ultra themes The "No clients yet" empty state had a hardcoded black color (rgba(0,0,0,0.45)) that vanished against the dark backgrounds. Drop the inline color, let it inherit from the AntD theme, and fade with opacity like the mobile card empty state already does. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): client-first tgbot add flow, tgId field, lightweight inbound options - tgbot: drop legacy per-protocol Add Client UI in favour of a client-first multi-inbound flow. New BuildClientDraftMessage / getInboundsAttachPicker let an admin pick one or more inbounds and submit a single client; per- protocol secrets are now generated server-side via fillProtocolDefaults. Drops awaiting_id/awaiting_password_tr/awaiting_password_sh state cases and add_client_ch_default_id/pass_tr/pass_sh/flow callbacks. Adds a setTGUser button + awaiting_tg_id state so the bot can set Client.TgID during Add. - clients UI: add Telegram user ID input to ClientFormModal (0 = none). Hide IP Limit field entirely when ipLimitEnable is off — disabled fields still take layout space, this collapses Auth(Hysteria) to full width. - inbounds API: new GET /panel/api/inbounds/options that returns just {id, remark, protocol, port, tlsFlowCapable}. Used by the clients page pickers so the dropdown payload stays small on panels with thousands of clients (drops settings JSON, clientStats, streamSettings). Server-side TlsFlowCapable mirrors Inbound.canEnableTlsFlow so the modal no longer needs to parse streamSettings client-side. - clientInfoMsg now shows attached inbound remarks, and getInboundUsages reports the attached client count per inbound. - api-docs: document the new /options endpoint and add tgId / flow to the clients add/update bodies. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(inbounds): keep Node column visible for node-attached inbounds The Node column was bound to hasActiveNode, so disabling every node hid the column even when inbounds were still attached to those nodes — the admin lost the visual cue that those inbounds belonged to a node and would come back when it was re-enabled. Combine hasActiveNode with a new hasNodeAttachedInbound check (any dbInbound with nodeId != null) so the column survives node-disable. * fix(api-docs): accept functional-component icons in EndpointSection AntD-Vue icons (SafetyCertificateOutlined, etc.) are functional components, so the icon prop's type: Object validator was rejecting them with a "Expected Object, got Function" warning at runtime. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * test: cover crypto, random, netsafe, sub helpers, xray equals, websocket hub, node service Adds ~110 unit tests across previously untested packages. Focus on pure-logic and concurrency surfaces where regressions would silently affect users: - util/crypto, util/random: password hashing round-trip, ss2022 key generation, alphabet/length invariants. - util/netsafe: IsBlockedIP edge cases, NormalizeHost validation, SSRF guard with AllowPrivate context bypass. - util/common, util/json_util: traffic formatter, Combine nil-skip, RawMessage empty-as-null and copy-on-unmarshal. - sub: splitLinkLines, searchKey/searchHost, kcp share fields, finalmask normalization, buildVmessLink round-trip. - xray: Config.Equals and InboundConfig.Equals field-by-field, getRequiredUserString/getOptionalUserString type checks. - web/websocket: hub registration, throttling, slow-client eviction, nil-receiver safety, concurrent register/unregister. - web/service: NodeService.normalize validation, normalizeBasePath, HeartbeatPatch.ToUI mapping. - web/job: atomicBool concurrent set/takeAndReset semantics. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * i18n(clients): replace English fallbacks with proper translation keys Pulls every hard-coded English label/title in the Clients page and its four modals through the i18n layer so localized panels stop leaking English. New keys live under pages.clients (auth, hysteriaAuth, uuid, flow, flowNone, reverseTag, reverseTagPlaceholder, telegramId, telegramIdPlaceholder, created, updated, ipLimit) plus refresh at the root and toasts.bulkDeletedMixed / bulkCreatedMixed for partial-failure toasts. Also switches the add-client modal's primary button from "Add" to "Create" for consistency with other create flows. The bulk-add Random/Random+Prefix/... email-method options stay hard-coded by request - they're identifier-shaped strings. * i18n: backfill 99 missing keys across all 12 non-English locales Brings every translation file up to parity with en-US.json so the Clients page, the fallback-children inbound section, the new refresh verb, the Nodes panel-version label and a handful of older holes stop falling through to the English fallback. New strings span: - pages.clients.* (labels, confirmations, toasts, emailMethods) - pages.inbounds.portFallback.* (Reality fallback inbound section) - pages.nodes.panelVersion, menu.clients, refresh Technical identifiers (Auth, UUID, Flow, Reverse tag) are intentionally left untranslated since they correspond to xray-core field names. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * i18n: drop stale pages.client block duplicated in every non-English locale Every non-English locale carried a pages.client (singular) section with 30 entries that duplicated pages.clients (plural). The plural namespace is what the Vue code actually consumes; the singular one was dead weight from an older rename that never got cleaned up in the non-English files. Removing it brings every locale to exactly 984 keys, matching en-US.json. * chore: apply modernize analyzer fixes across codebase Mechanical replacements suggested by golang.org/x/tools/.../modernize: strings.Cut/CutPrefix/SplitSeq, slices.Contains, maps.Copy, min(), range-over-int, new(expr), strings.Builder for hot += loops, reflect.TypeFor[T](), sync.WaitGroup.Go(), drop legacy //+build lines. * feat(database): add PostgreSQL as an optional backend alongside SQLite Lets operators with large client counts or multi-node setups pick PostgreSQL at install time without breaking the existing SQLite default. Backend is selected at runtime via XUI_DB_TYPE/XUI_DB_DSN, a small dialect layer keeps the five JSON_EXTRACT/JSON_EACH queries portable, and a new `x-ui migrate-db` subcommand copies SQLite data into PostgreSQL in FK-aware order. * fix(inbounds): gate node selector to multi-node-capable protocols Hide the Deploy-To selector and clear nodeId when switching to a protocol that can't run on a remote node. Also: - subs: return 404 (not 400) when subId matches no inbounds, so VPN clients distinguish "deleted/unknown" from a server error - hysteria link gen: use the inbound's resolved address so node-managed inbounds advertise the node host instead of the central panel - shadowsocks: default network to 'tcp' (udp was causing issues for some clients on first-create) - vite dev proxy: rewrite migrated-route bypass against the live base path instead of a hardcoded single-segment regex Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): bulk add/delete correctness + perf, working pagination, delayed-start in form Bulk add/delete were serial on the frontend (one toast per call, N round-trips) and the backend race exposed by parallelizing them lost client attachments and hit UNIQUE constraint failed on client_inbounds. The single add/edit modal also had no Start-After-First-Use option, and the table never showed the delayed duration. Backend (web/service/client.go): - Per-inbound mutex on Add/Update/Del InboundClient so concurrent writers on the same inbound don't lose the read-modify-write of settings JSON. - SyncInbound skips create+join when the email is tombstoned so a concurrent maintenance pass (adjustTraffics, autoRenewClients, markClientsDisabledIn- Settings) that did a stale RMW can't resurrect a just-deleted client with a fresh id. - compactOrphans sweeps settings.clients entries whose ClientRecord no longer exists, applied in Add/DelInboundClient + DelInboundClientByEmail so each user-initiated mutation self-heals the inbound's settings. - DelInboundClient uses Pluck instead of First for the stats lookup so a missing row doesn't abort the delete with a noisy ErrRecordNotFound log. Frontend: - HttpUtil.{get,post} accept a silent option that suppresses the auto-toast. - ClientBulkAddModal fires creates in parallel + silent + one summary toast. - useClients.removeMany runs deletes in parallel + silent and refreshes once; ClientsPage bulk delete uses it and shows one aggregate toast. - useClients.applyInvalidate debounces 200 ms so the burst of N WebSocket invalidate events from the backend collapses into a single refresh. - ClientsPage pagination is reactive (paginationState ref + tablePagination computed); onTableChange persists page-size and page changes. - ClientFormModal gains a Start-After-First-Use switch + Duration days input alongside the existing Expiry Date picker; on edit-mode open a negative expiryTime is decoded back to delayed mode + days; on submit the payload sends -86400000 * days or the absolute timestamp. - ClientsPage table shows the delayed-start duration (blue tag Nd, tooltip Start After First Use: Nd) instead of infinity. - Telegram ID field in the form is hidden when /panel/setting/defaultSettings reports tgBotEnable=false; Comment then fills the row. - Form row 3 collapses UUID (span 12) + Total GB (span 8) + Limit IP (span 4) when ipLimitEnable is on, else UUID + Total GB at 12/12. - useInbounds.rollupClients counts only clients with a matching clientStats row, so orphans in settings.clients no longer inflate the inbound's count. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(windows): clean shutdown, working panel restart, harden kernel32 load Three Windows-specific issues addressed: 1. Orphaned xray-windows-amd64 after VS Code debugger stop. Delve's "Stop" sends TerminateProcess to the Go binary, which is uncatchable — our signal handlers never run, so xrayService.StopXray() is skipped and xray is left dangling. Spawn xray as a child of a Job Object with JOB_OBJECT_LIMIT_KILL_ON_JOB_CLOSE so the OS kills xray when our handle to the job is closed (which happens even on TerminateProcess). Also trap os.Interrupt in main so Ctrl+C in the terminal runs the graceful path. 2. /panel/setting/restartPanel logged "failed to send SIGHUP signal: not supported by windows" because Windows can't deliver arbitrary signals. Add a restart hook in web/global; main registers it to push SIGHUP into its own signal channel, and RestartPanel calls the hook before falling back to the (Unix-only) signal path. Same restart-loop code runs in both cases. 3. util/sys/sys_windows.go now uses windows.NewLazySystemDLL so the kernel32.dll resolve is pinned to %SystemRoot%\System32 (prevents DLL hijacking by a planted DLL next to the binary). Local filetime type replaced with windows.Filetime, and the unreliable syscall.GetLastError() fallback replaced with a type assertion on the errno captured at call time. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(sys): correct CPU/connection accounting on linux + darwin util/sys/sys_linux.go: - GetTCPCount/GetUDPCount were counting the column header row in /proc/net/{tcp,udp}[6] as a connection, inflating the reported total by 1 per non-empty file (so the panel status line always showed 2 more connections than actually existed). Replace getLinesNum + safeGetLinesNum with a single bufio.Scanner-based countConnections that skips the header. - CPUPercentRaw now opens HostProc("stat") instead of a hardcoded /proc/stat so HOST_PROC overrides apply, matching the connection counters in the same file. - Simplify CPU field unpacking: pad nums to 8 once instead of guarding every assignment with a len check. util/sys/sys_darwin.go: - Fix swapped idle/intr indices on kern.cp_time. BSD CPUSTATES order is user, nice, sys, intr, idle (CP_INTR=3, CP_IDLE=4) — gopsutil's cpu_darwin_nocgo.go reads the same layout. The previous code used out[3] as idle and out[4] as intr, so busy = total - dIdle was actually subtracting interrupt time, making the panel report CPU usage close to 100% on macOS regardless of actual load. - Collapse the per-field delta math into a single loop. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(xray): rotate crash reports into log folder, prevent overwrites writeCrashReport had two flaws: it wrote to the bin folder (alongside the xray binary) which conflates artifacts, and the second-precision timestamp meant a tight restart-loop crash burst overwrote prior reports. Write to the log folder with nanosecond precision and keep the last 10 reports. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * revert(inbounds): drop unreleased portfallback protocol The Port-with-Fallback inbound (commit 62fd9f9d) was confusing as a standalone protocol — fallbacks belong on a regular VLESS/Trojan TCP-TLS inbound, the way Xray models them natively. Rip out the entire feature cleanly (no migration needed since it was never released): protocol constant, fallback children DB table, FallbackService, 2 API endpoints, all UI rows, related translations and api-docs. A native fallback flow attached to VLESS/Trojan TCP-TLS/Reality will land in a follow-up commit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds): native fallbacks on VLESS/Trojan TCP-TLS, with working child links A VLESS or Trojan inbound on TCP with TLS or Reality can now act as a fallback master: pick existing inbounds as children and the panel auto- fills the SNI / ALPN / path / xver routing fields from each child's transport, auto-builds settings.fallbacks at config-gen time, and rewrites the child's client-share link so it advertises the master's reachable endpoint and TLS state instead of the child's loopback listen. Layout matches the Xray All-in-One Nginx example: master at :443 with clients + TLS, each child on 127.0.0.1 with its own transport+clients. Order matters (Xray walks fallbacks top-to-bottom) — reorder via the per-row up/down arrows. Path / SNI / ALPN are exposed under a per-row Edit toggle for the rare cases where the auto-derivation needs overriding; otherwise just pick a child and you're done. Backend: new InboundFallback table + FallbackService (GetByMaster / SetByMaster / GetParentForChild / BuildFallbacksJSON); two routes (GET / POST /panel/api/inbounds/:id/fallbacks); xray.GetXrayConfig injects settings.fallbacks for any VLESS/Trojan TCP-TLS/Reality inbound; GetInbounds annotates each child with FallbackParent so the frontend can rewrite links without an extra round-trip. Link projection covers every emission path — clients-page QR/links, per-inbound Get URL, raw subscription, sub-JSON, sub-Clash, and the inbounds-page link/info/QR — via a shared projectThroughFallbackMaster on the backend and a shared projectChildThroughMaster on the frontend that both handle the panel-tracked relationship and the legacy unix-socket (@vless-ws) convention. Strings translated into all 12 non-English locales. * docs: rewrite CONTRIBUTING with full local-dev setup The prior three-line CONTRIBUTING left newcomers guessing at every non-trivial step: which Go / Node versions, where xray comes from, why the panel goes blank when XUI_DEBUG=true is flipped on, how the Vue multi-page setup is wired, what to do on Windows when go build trips on the CGo SQLite driver. Now covers prerequisites, MinGW-w64 install on Windows (niXman builds or MSYS2), one-shot first-time setup, two frontend dev workflows with the XUI_DEBUG asset-cache gotcha called out, the architecture and conventions of the Vue side, a project-layout map, useful env vars, and the PR checklist. ---------
2026-05-19 03:16:42 -07:00
else
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
read -rp "Would you like to customize the Panel Port settings? (If not, a random port will be applied) [y/n]: " config_confirm
if [[ "${config_confirm}" == "y" || "${config_confirm}" == "Y" ]]; then
read -rp "Please set up the panel port: " config_port
echo -e "${yellow}Your Panel Port is: ${config_port}${plain}"
else
config_port=$(shuf -i 1024-62000 -n 1)
echo -e "${yellow}Generated random port: ${config_port}${plain}"
fi
2026-05-19 04:09:35 -07:00
fi
${xui_folder}/x-ui setting -username "${config_username}" -password "${config_password}" -port "${config_port}" -webBasePath "${config_webBasePath}"
echo ""
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${green} SSL Certificate Setup (RECOMMENDED) ${plain}"
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${yellow}SSL is strongly recommended. Skip only if a reverse proxy${plain}"
echo -e "${yellow}or SSH tunnel handles TLS for you.${plain}"
echo -e "${yellow}Let's Encrypt now supports both domains and IP addresses!${plain}"
echo ""
prompt_and_setup_ssl "${config_port}" "${config_webBasePath}" "${server_ip}"
# Retrieve the API token for display
fix(cli): let -getApiToken name the token it regenerates (#6405) * fix(cli): let -getApiToken name the token it regenerates The flag's help text said "Display current API token". It cannot display anything -- tokens are stored as SHA-256 hashes, and the command's own first two output lines say so. What it does is destroy and reissue a credential: GetApiToken calls RecreateByName on the hardcoded name "cli-fallback". The help therefore invited an operator to run a command they believed was read-only, and it revoked a token someone else was holding. Because that name is a single global slot, two callers silently invalidate each other, and the loser is left with a token that answers HTTP 404 with an empty body -- indistinguishable from a wrong base path, so the failure does not even say what happened. install.sh is one of those callers, at lines 1231 and 1325, so the collision already exists inside this repository. Add -tokenName, defaulting to cli-fallback so install.sh and every existing invocation behave exactly as before. -getApiToken stays a boolean on purpose: install.sh calls it as `x-ui setting -getApiToken true`, and a string flag would swallow that trailing argument and mint a token named "true". The name now reaches both branches of GetApiToken. On a database with no tokens the command used to create one called "install", which the CLI could then never rotate -- defeating the stated purpose of the cli-fallback constant, that -getApiToken cannot accumulate admin-equivalent credentials it never revokes. Both branches use the resolved name, so repeated calls rotate a single slot instead of leaving a permanent token behind. Also cap the name at 64 characters in RecreateByName. Create already enforces that limit on the same column; RecreateByName did not, and it now receives operator input. Assisted-by: Claude Code:claude-opus-5 (mostly) * fix(cli): keep the installer's token out of the rotated slot Folding both branches of GetApiToken onto one name made the bug worse in the exact case this change is about. install.sh records the token it gets on a fresh panel; with both branches on cli-fallback, the next bare -getApiToken rotated that very row and silently invalidated the credential written into the install-result file. Restore the split default -- "install" when the database has no tokens, cli-fallback when it does -- so nothing about an unnamed call changes. An explicit -tokenName still applies to both branches, which is what keeps the flag coherent: -tokenName ci-bot now yields ci-bot on a fresh panel too, rather than "install". Pin it with a test that reads the install row's id and hash before and after a rotation, since a name-only assertion would pass against a deleted-and- recreated row. * fix(cli): stop the `-getApiToken true` form from swallowing -tokenName Three corrections from review. Go's flag package stops parsing at the first non-flag argument, so the trailing `true` in install.sh's invocation does not merely get ignored -- it terminates parsing. An operator copying that documented shape and writing `x-ui setting -getApiToken true -tokenName ci-bot` left tokenName empty, so the command rotated cli-fallback: the shared-slot collision this change exists to remove, reachable through the one form the repository itself demonstrates. Verified against the built binary, which printed `The API token "cli-fallback" has been regenerated`. Drop the stray `true` from both install.sh call sites so the documented form no longer teaches the trap, and warn whenever `setting` is given positional arguments, naming what was ignored. A warning rather than an error, because an older install.sh in the wild still passes `true` and must keep working. Cover both branches in the help strings. They described only the rotation path, so on a fresh panel -h announced cli-fallback while the command actually mints `install`, and nothing is regenerated or invalidated there at all -- misleading help being the defect this change set out to remove. Assert the concrete error in the name-length test. It checked only that some error came back, which RecreateByName's empty-name guard and its transaction errors would satisfy just as well.
2026-09-08 05:15:37 -07:00
local config_apiToken=$(${xui_folder}/x-ui setting -getApiToken | grep -Eo 'apiToken: .+' | awk '{print $2}')
2026-05-19 04:09:35 -07:00
# Display final credentials and access information
echo ""
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${green} Panel Installation Complete! ${plain}"
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${green}Username: ${config_username}${plain}"
echo -e "${green}Password: ${config_password}${plain}"
echo -e "${green}Port: ${config_port}${plain}"
echo -e "${green}WebBasePath: ${config_webBasePath}${plain}"
echo -e "${green}Database: ${db_label}${plain}"
2026-05-19 04:09:35 -07:00
echo -e "${green}Access URL: ${SSL_SCHEME}://${SSL_HOST}:${config_port}/${config_webBasePath}${plain}"
echo -e "${green}API Token: ${config_apiToken}${plain}"
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${yellow}⚠ IMPORTANT: Save these credentials securely!${plain}"
if [[ "$SSL_SCHEME" == "https" ]]; then
echo -e "${yellow}⚠ SSL Certificate: Enabled and configured${plain}"
else
echo -e "${yellow}⚠ SSL Certificate: Skipped — panel is HTTP-only. Use a reverse proxy or SSH tunnel.${plain}"
Feat/multi inbound clients (#4469) * feat(clients): add shadow tables for first-class client promotion Introduces three new GORM-backed tables (clients, client_inbounds, inbound_fallback_children) and a populate-only seeder that backfills them from each inbound's existing settings.clients JSON. Duplicate emails across inbounds auto-merge under one client row, with each field conflict logged. Existing services are unchanged and continue reading from settings.clients — this commit is groundwork only. * feat(clients): make clients+client_inbounds the runtime source of truth Adds ClientService.SyncInbound that reconciles the new tables from each inbound's clients list whenever existing service paths mutate settings.clients. Wires it into AddInbound, UpdateInbound, AddInboundClient, UpdateInboundClient, DelInboundClient, DelInboundClientByEmail, DelDepletedClients, autoRenewClients, and the timestamp-backfill path in adjustTraffics, plus DetachInbound on DelInbound. GetXrayConfig now builds settings.clients from the new tables before writing config.json, and getInboundsBySubId joins through them instead of JSON_EACH on settings JSON. Live Xray config and subscription endpoints are now driven by the relational view; settings.clients JSON stays in step as a side effect of every write. * feat(clients): add top-level Clients tab and CRUD API Adds /panel/api/clients endpoints (list, get, add, update, del, attach, detach) backed by ClientService methods that orchestrate the per-inbound Add/Update/Del flows so a single client row is created once and attached to many inbounds in one operation. The frontend gains a dedicated Clients page (frontend/clients.html + src/pages/clients/) with an AntD table, multi-inbound attach modal, and full CRUD. Axios interceptor learns to honour Content-Type: application/json so the JSON endpoints work alongside the legacy form-encoded ones. The legacy per-inbound client modal stays untouched in this PR — both flows now write to the same source of truth. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds): add Port-with-Fallback inbound type Adds a new "portfallback" protocol that emits as a VLESS-TLS inbound under the hood but is paired with a sidecar table of child inbounds. Panel auto-builds settings.fallbacks at Xray-config-gen time from the sidecar — each child's listen+port becomes the fallback dest, with SNI/ALPN/path/xver match criteria pulled from the row. No more typing loopback ports by hand or keeping settings.fallbacks in sync. Backend: new FallbackService (Get/SetChildren, BuildFallbacksJSON); two new routes (GET/POST /panel/api/inbounds/:id/fallbackChildren); xray.GetXrayConfig injects fallbacks for PortFallback inbounds; the inbound model emits protocol="vless" so Xray accepts the config. Frontend: PORTFALLBACK joins the protocol dropdown; selecting it shows the standard VLESS controls plus a Fallback Children table (inbound picker + per-row SNI/ALPN/path/xver). Children are loaded on edit and replaced atomically on save. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Reset Traffic, QR Code, Info actions + Online/Remaining columns The Clients page table gains: - Online column — green/grey tag driven by /panel/api/inbounds/onlines, polled every 10s. - Remaining column — bytes-remaining tag, coloured green/orange/red against quota, purple infinity when unlimited. - Action icons per row: QR, Info, Reset traffic, Edit, Delete. ClientInfoModal shows the full client detail (uuid/password/auth, traffic ↑/↓ + remaining + all-time, expiry absolute + relative, attached inbounds chip list, online + last-online). ClientQrModal fetches links for the client's subId via /panel/api/inbounds/getSubLinks/:subId and renders each one through the existing QrPanel component. Reset Traffic confirms then calls the existing per-inbound endpoint on the client's first attached inbound (the traffic row is keyed on email globally, so any attached inbound resets the shared counter). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): expose Attached inbounds in edit mode The multi-select was gated on add-only, so editing a client had no way to change which inbounds it belonged to. The picker now shows in both modes, and on submit the modal diffs the picked set against the original attachedIds — additions go through the /attach endpoint, removals through /detach, both after the field update lands so the new attachments get the latest values. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): unbreak template parsing + stale i18n keys - InboundFormModal: split the multi-line help string in the PortFallback section onto one line — Vue's template parser was bailing on Unterminated string constant because a single-quoted literal spanned two lines inside a {{ }} interpolation. - ClientInfoModal: t('disable') was missing at the root level, so vue-i18n returned the key path literally. Use t('disabled') which exists. - Linter cleanup elsewhere: pages.client.* references renamed to pages.clients.* to match the merged i18n block; whitespace normalisation in a few unrelated Vue templates. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * 1 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(traffic): drop all-time traffic tracking Removes the AllTime field from Inbound and ClientTraffic and migrates existing DBs by dropping the all_time columns on startup. The counter duplicated up+down without adding signal, and the per-event accumulator ran on every traffic write. Frontend: drop the All-time column from the inbound list and the client-row table, the All-time row from the client info modal, and the All-Time Total Usage tile from the inbounds summary card. The allTimeTraffic/allTimeTrafficUsage i18n keys are removed across every locale. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): mobile cards, multi-select, bulk add Adds the same row-card layout the inbounds page uses on mobile: the table is suppressed under the mobile breakpoint and each client renders as a compact card with a status dot, email, Info button, Enable switch, and overflow menu. All the per-client detail (traffic, remaining, expiry, attached inbounds, flow, created/updated, URL, subscription) opens through the existing info modal. Multi-select with bulk delete wires AntD row-selection on desktop and a per-card checkbox on mobile; a Delete (N) button appears in the toolbar when anything is selected. Bulk add reuses the five email-generation modes from the inbound bulk modal but takes a multi-inbound picker so one bulk run can attach to several inbounds at once. Submits client-by-client through the existing /panel/api/clients/add endpoint. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(inbounds): remove legacy per-inbound client UI Now that clients live as first-class rows attached to one or many inbounds, the per-inbound client UI on the inbounds page is dead weight — every client action either has a global equivalent on the Clients page or makes no sense in a many-to-many world. Deletes ClientFormModal, ClientBulkModal, CopyClientsModal, and ClientRowTable from inbounds/. Strips the matching emits, refs, handlers, and dropdown menu items from InboundList and InboundsPage, and removes the dead mobile expand-chevron state and the desktop expanded-row plumbing that drove the inline client table. The InboundFormModal Clients tab still works in add-mode (one inline client at inbound creation) — that flow goes through ClientService. SyncInbound on save and remains useful. Fixes a stray "</a-dropdown>" left over by an earlier toolbar edit in ClientsPage that broke the template parser. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Delete depleted action Mirrors the legacy delDepletedClients action that lived under the inbounds page, but as a first-class /panel/api/clients/delDepleted endpoint backed by ClientService. The new path goes through ClientService.Delete for each depleted email, so the new clients + client_inbounds + xray_client_traffic tables stay consistent. Adds a danger-styled toolbar button on the Clients page (next to Reset all client traffic) with a confirm dialog and a toast reporting the deleted count. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(api): move every client-shaped endpoint off /inbounds onto /clients After the multi-inbound client migration, client state belongs to the client API surface, not the inbound one. Twelve routes that were crammed under /panel/api/inbounds/* now live where they belong, under /panel/api/clients/*. Moved (route, handler, doc): POST /clientIps/:email POST /clearClientIps/:email POST /onlines POST /lastOnline POST /updateClientTraffic/:email POST /resetAllClientTraffics/:id POST /delDepletedClients/:id POST /:id/resetClientTraffic/:email GET /getClientTraffics/:email GET /getClientTrafficsById/:id GET /getSubLinks/:subId GET /getClientLinks/:id/:email Their /clients/* counterparts are: POST /clients/clientIps/:email POST /clients/clearClientIps/:email POST /clients/onlines POST /clients/lastOnline POST /clients/updateTraffic/:email POST /clients/resetTraffic/:email (email-only, fans out) GET /clients/traffic/:email GET /clients/traffic/byId/:id GET /clients/subLinks/:subId GET /clients/links/:id/:email per-inbound resetAllClientTraffics and delDepletedClients are dropped entirely — the Clients page already exposes global Reset All Traffic and Delete depleted actions, and per-inbound resets are meaningless once a client can be attached to many inbounds. ClientService.ResetTrafficByEmail is the new email-only reset path: it looks up every inbound the client is attached to and pushes the counter reset + Xray re-add through inboundService.ResetClientTraffic for each one, so depleted users come back online instantly. Frontend callers (ClientsPage, useClients, ClientQrModal, ClientInfoModal, InboundInfoModal, InboundsPage, useInbounds) all switched to the new paths. The Inbounds page drops its per-inbound "Reset client traffic" and "Delete depleted clients" dropdown items — users do those at the client level now. api-docs is rebuilt to match. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(service): switch tgbot + ldap callers to ClientService Adds two thin helpers to ClientService (CreateOne, DetachByEmail) and rewrites tgbot.SubmitAddClient and ldap_sync_job to call ClientService directly. Removes the JSON-blob payloads (BuildJSONForProtocol output for add, clientsToJSON/clientToJSON helpers) that callers previously fed to InboundService.AddInboundClient/DelInboundClient. ldap_sync_job.batchSetEnable now loops InboundService.SetClientEnableByEmail per email instead of trying to coerce AddInboundClient into doing the update — the old path would have failed duplicate-email validation for existing clients anyway. The legacy InboundService.AddInboundClient/UpdateInboundClient/ DelInboundClient methods stay in place; they are now only used internally by ClientService Create/Update/Delete/Attach. Inlining + deleting them follows in a separate commit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(service): move all client mutation methods to ClientService Moves the client mutation surface out of InboundService and into ClientService. These methods all operate on a single client (identity fields, traffic limits, expiry, ip limit, enable state, telegram tg id) and didn't belong on the inbound aggregate. Moved (12 methods): AddInboundClient, UpdateInboundClient, DelInboundClient, DelInboundClientByEmail, checkEmailsExistForClients, SetClientTelegramUserID, checkIsEnabledByEmail, ToggleClientEnableByEmail, SetClientEnableByEmail, ResetClientIpLimitByEmail, ResetClientExpiryTimeByEmail, ResetClientTrafficLimitByEmail. Each method now takes an explicit *InboundService for the helpers that legitimately stay on InboundService (GetInbound, GetClients, runtimeFor, AddClientStat / UpdateClientStat / DelClientStat, DelClientIPs / UpdateClientIPs, emailUsedByOtherInbounds, getAllEmailSubIDs, GetClientInboundByEmail / GetClientInboundByTrafficID, GetClientTrafficByEmail). Stays on InboundService: ResetClientTrafficByEmail and ResetClientTraffic(id, email) — these mutate xray_client_traffic rows, not client identity, so they're inbound-side bookkeeping. Callers updated: tgbot (6 calls), ldap_sync_job (1 call), InboundService internal (writeBackClientSubID, CopyInboundClients, AddInbound's email-uniqueness check), ClientService Create/Update/ Delete/Attach/Detach. Also removes a dead resetAllClientTraffics controller handler whose route was already gone after the previous /clients API migration. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(clients): finish migrating to ClientService + tidy IP routes Two related cleanups in the new /clients surface: 1. Move ResetAllClientTraffics (bulk-reset of xray_client_traffic + last_traffic_reset_time, with node-runtime propagation) from InboundService to ClientService. PeriodicTrafficResetJob now holds a clientService and calls j.clientService.ResetAllClientTraffics(&j.inboundService, id). The last client-mutation method on InboundService is gone. 2. Shorten redundantly-named routes/handlers under /panel/api/clients: - /clientIps/:email -> /ips/:email (handler getIps) - /clearClientIps/:email -> /clearIps/:email (handler clearIps) The "client" prefix was redundant inside the clients namespace. Frontend (InboundInfoModal) and api-docs updated to match. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds,clients): clean up inbound modal + enrich client modal Inbound modal rework (InboundFormModal.vue + inbound.js): - Drop the embedded Client subform in the Protocol tab. Multi-inbound clients are managed exclusively from the Clients page now; a fresh inbound is created with zero clients (settings constructors default to []) and the user attaches clients afterwards. - Hide the Protocol tab entirely when it has nothing to render (VMESS, Trojan without fallbacks, Hysteria). Auto-switches active tab to Basic when the tab disappears while focused. - Move the Security section (Security selector + TLS block with certs and ECH + Reality block) out of the Stream tab into its own Security tab, sharing the canEnableStream gate. Client modal additions (ClientFormModal.vue + ClientBulkAddModal.vue): - Flow select (xtls-rprx-vision / -udp443) appears only when the panel actually has a Vision-capable inbound (VLESS or PortFallback on TCP with TLS or Reality). Hidden otherwise, and cleared when it disappears. - IP Limit input is disabled when the panel-level ipLimitEnable setting is off, fetched into useClients alongside subSettings and threaded through ClientsPage to both modals. - Edit modal now shows an "IP Log" section listing IPs that have connected with the client's credentials, with refresh and clear buttons (calls the renamed /panel/api/clients/ips and /clearIps endpoints). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(inbounds): drop manual Fallbacks UI from inbound modal The PortFallback protocol type now covers the common VLESS-master-plus-children case with auto-wired dests, so the manual Fallbacks editor (showFallbacks block in the Protocol tab) is mostly redundant. Removed: - the v-if="showFallbacks" template block (SNI/ALPN/Path/dest/PROXY rows) - the showFallbacks computed - the addFallback / delFallback helpers - the .fallbacks-header / .fallbacks-title styles - the showFallbacks gate from hasProtocolTabContent (so Trojan-over-TCP no longer shows an empty Protocol tab) Power users who need a non-inbound fallback dest (nginx, static site) can still author settings.fallbacks via the Advanced JSON tab. * feat(clients,inbounds): move search/filter to Clients page + small fixes Search/filter relocation: - Remove the search/filter toolbar (search switch + filter radio + protocol/node selects + the visibleInbounds projection + inboundsFilterState localStorage + filter CSS + the SearchOutlined/ FilterOutlined/ObjectUtil/Inbound imports it required) from InboundList. The filters were all client-oriented buckets bolted onto the inbound row. - Add a search/filter toolbar to ClientsPage with the same shape: switch between deep-text search and bucket filter (active / deactive / depleted / expiring / online) + protocol filter that matches clients attached to at least one inbound with the chosen protocol. State persists in clientsFilterState localStorage. filteredClients drives both the desktop table and the mobile card list, and select-all / allSelected / someSelected only span the visible subset. - useClients now also fetches expireDiff and trafficDiff from /panel/setting/defaultSettings (used to detect the expiring bucket); ClientsPage threads them into the client-bucket helper. Loose fixes folded in: - Add Client: email field is auto-filled with a random handle on open, matching uuid/subId/password/auth. - Inbound clone: parse and reuse the source settings JSON (with clients reset to []) instead of building a fresh defaulted Settings, so VLESS Encryption/Decryption and other non-client fields survive the clone. - en-US.json: add the ipLog string used by the edit-client modal. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Reverse tag field for VLESS-attached clients Mirrors the Flow field's pattern: a Reverse tag input appears in the Add/Edit Client modal whenever at least one selected inbound is VLESS or PortFallback. The value rides over the wire as client.reverse = { tag: '...' } so it lands directly in model.Client's *ClientReverse field; an empty value omits the reverse key entirely. On edit the field is hydrated from props.client.reverse?.tag, and the showReverseTag watcher clears the field if the user drops the last VLESS-like inbound from the selection. * fix(xray): emit only protocol-relevant fields per client entry The Xray config synthesizer was writing every identifier field (id, password, flow, auth, security/method, reverse) on every client entry regardless of the inbound's protocol. Xray ignores unknown fields, so the config worked, but it diverged from the spec and leaked secrets across protocols when one client was attached to multiple inbounds — a VLESS inbound's generated config carried the same client's Trojan password and Hysteria auth alongside its uuid. Switch on inbound.Protocol when building each entry: - VLESS / PortFallback: id, flow, reverse - VMess: id, security - Trojan: password, flow - Shadowsocks: password, method - Hysteria / Hysteria2: auth email is emitted for every protocol. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): restore auto-disable kick under new schema disableInvalidClients still resolved (inbound_tag, email) pairs via JSON_EACH(inbounds.settings.clients), which is empty after migrating to the clients + client_inbounds tables. Result: xrayApi.RemoveUser never ran for depleted clients, clients.enable stayed true so the UI showed them as active, and only xray_client_traffic.enable got flipped - making "Restart Xray After Auto Disable" only half-work. Resolve the targets via a JOIN through the new schema, flip clients.enable so the Clients page reflects the state, and drop the legacy JSON write-back plus the subId cascade workaround (email is unique now). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): live WebSocket updates + Ended status surfacing ClientsPage now subscribes to traffic / client_stats / invalidate WebSocket events instead of polling /onlines every 10s. Per-row traffic counters refresh in place, online state stays current, and list-level mutations elsewhere trigger a refresh. The client roll-up summary moves from InboundsPage to ClientsPage where it belongs, restructured into six labeled stat tiles (Total / Online / Ended / Expiring / Disabled / Active) with email popovers on the ones with issues. Auto-disabled clients (traffic exhausted or expiry passed) now classify as 'depleted' even though clients.enable=false, so they show up under the Ended filter and render a red Ended tag instead of looking indistinguishable from an operator-disabled row. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(nodes): per-node client roll-up and panel version Added transient inboundCount / clientCount / onlineCount / depletedCount fields to model.Node, populated by NodeService.GetAll via aggregated queries (one join across inbounds + client_inbounds, one over client_traffics intersected with the in-memory online emails). The Nodes list renders these as colored chips on a new "Clients" column so an operator can see at a glance how many users each node carries and how many are currently online or depleted. Also exposes the remote panel's version. The central panel adds panelVersion to its /api/server/status payload (sourced from config.GetVersion). Probe reads that field and persists it on the node row, mirroring how xrayVersion already flows. NodesPage gets a new column next to Xray Version, in both desktop and mobile views, with English and Persian strings. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): stop node sync from resurrecting deleted clients Several related issues around node-managed clients: - Remote runtime: drop the per-inbound resetAllClientTraffics path and point traffic/onlines/lastOnline fetches at the new /panel/api/clients/* routes. - Delete from master: always push the updated inbound to the node even when the client was already disabled or depleted, so the node actually loses the user instead of silently keeping it. - setRemoteTraffic: mirror remote clients into the central tables only on first discovery of a node inbound. Matched inbounds let the master own the join table, so a stale snap can no longer re-create a ClientRecord (and join row) for a client that was just deleted on the master. - ClientService.Delete: route through submitTrafficWrite so deletes serialize with node traffic merges, and switch the final ClientRecord delete to an explicit Where("id = ?") clause. - setRemoteTraffic UNIQUE-constraint fix: use clause.OnConflict on inserts and email-keyed UPDATEs for client_traffics, so mirroring a snap doesn't trip the unique email index. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(clients): switch client API endpoints from id to email All client-scoped routes now use the unique email as the path key (get, update, del, attach, detach, links). Email is the stable, protocol-independent identifier — UUIDs don't exist for trojan or shadowsocks, and internal numeric ids leaked panel implementation detail into the public API. Removed the redundant /traffic/byId/:id endpoint (covered by /traffic/:email) and collapsed /links/:id/:email into /links/:email, which now returns links across every attached inbound for the client. Frontend selection, bulk delete, and toggle state are now keyed by email as well, dropping the id→email lookup workaround. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(server): move cached state and helpers into ServerService ServerController had grown to hold its own status cache, version-list TTL cache, history-bucket whitelist, and the loop that drove all three — concerns that belong in the service layer. Pull them out: - lastStatus + the @2s refresh become ServerService.RefreshStatus and ServerService.LastStatus; the controller's cron now just orchestrates the cross-service side effects (xrayMetrics sample, websocket broadcast). - The 15-minute Xray-versions cache (with stale-on-error fallback) moves into ServerService.GetXrayVersionsCached, collapsing the controller handler to a single call. - The freedom/blackhole outbound-tag walk used by /xraylogs becomes ServerService.GetDefaultLogOutboundTags. - The allowed-history-bucket whitelist moves to package-level service.IsAllowedHistoryBucket, so both NodeController and ServerController validate against the same list. Net result: web/controller/server.go drops from 458 to 365 lines and contains only HTTP wiring + presentation-y side effects. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(api): emit JSON-text columns as nested objects Inbound, ClientRecord, and InboundClientIps store settings / streamSettings / sniffing / reverse / ips as JSON-text in the DB. The API was passing that text through verbatim, so every consumer had to JSON.parse a string inside a string. Add MarshalJSON / UnmarshalJSON so the wire format is a real nested object, while still accepting the legacy escaped-string shape on write. Frontend dbinbound.js gets a matching coerceInboundJsonField helper for the same dual-shape read path, and inbound.js toJson stops emitting empty/placeholder fields (externalProxy [], sniffing destOverride when disabled, etc.) so the new normalised JSON stays terse. api-docs and the inbound-clone path are updated to the new shape. Controller route lists are regrouped so all GETs sit above POSTs. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): include inboundIds and traffic in /clients/list ClientRecord got its own MarshalJSON in the previous commit, and ClientWithAttachments embeds it to add inboundIds and traffic. Go promotes the embedded MarshalJSON to the outer struct, so the encoder was calling ClientRecord.MarshalJSON for the whole value and silently dropping the extras. The frontend reads row.inboundIds / row.traffic from /clients/list, so attached inbounds didn't render and newly added clients looked like they hadn't saved. Add an explicit MarshalJSON on ClientWithAttachments that splices the extras in. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): gate IP Log on ipLimitEnable + clean access-log dropdown Legacy panel hid the IP Log section when access logging was off; the Vue 3 migration left it gated on isEdit only, so the section showed even when xray's access log was 'none' and nothing was being recorded. Restore the ipLimitEnable gate on the edit modal's IP Log form-item. While here, clean up the Xray Settings access-log dropdown: previously two 'none' entries appeared (an empty value labelled with t('none') and the literal 'none' from the options array). Drop the empty option for access log (the literal 'none' covers it) and relabel the empty option for error log / mask address to t('empty') so they're distinguishable. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(nodes): route per-client ops through node clients API + orphan sweep Adds Runtime methods AddClient, UpdateUser, and DeleteUser so master mutates clients on a node via /panel/api/clients/{add,update,del} rather than pushing the whole inbound. The previous rt.UpdateInbound path made the node DelInbound+AddInbound on every single-client change, briefly cycling every other user on the same inbound. DelInbound no longer filters by enable=true, so a disabled node inbound actually gets removed from the node instead of being resurrected by the next snap. setRemoteTrafficLocked now sweeps any ClientRecord with zero ClientInbound rows after SyncInbound rebuilds the attachments, which is how a node-side delete propagates back to master instead of leaving a detached ghost. ClientService.Delete tombstones the email first so a snap arriving mid-delete can't re-create the record. WebSocket broadcasts an "invalidate(clients)" message on every client mutation so the Clients page refreshes without manual reload. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(balancers): allow fallback on all strategies + feed burstObservatory from random/roundRobin Drops the random/roundRobin gate on the Fallback field in BalancerFormModal so every strategy can pick a fallback outbound. syncObservatories now feeds burstObservatory from leastLoad + random + roundRobin balancers (was leastLoad only), matching how leastPing feeds observatory. Fix the JsonEditor "Unexpected end of JSON input" that appeared when switching a balancer between leastPing and another strategy: the obsView watcher was gated on showObsEditor (a boolean OR of the two flags) and missed the case where one observatory swapped for the other in the same tick. Watch the individual flags instead so obsView flips to the surviving editor and the getter stops pointing at a deleted key. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(inbounds): use sortedInbounds for mobile empty-state check InboundList referenced an undefined visibleInbounds in the mobile card list's empty-state guard, throwing "Cannot read properties of undefined (reading 'length')" and breaking the entire mobile render. * feat(clients): sortable table columns Adds the same sortState / sortableCol / sortFns pattern InboundList uses, wrapping filteredClients in sortedClients so sort composes with the existing search/filter pipeline. Sortable: enable, email, inboundIds (attachment count), traffic, remaining, expiryTime; actions and online stay unsorted. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(shadowsocks): generate valid ss2022 keys and per-client method for legacy ciphers The Add Client flow on shadowsocks inbounds was producing xray configs that failed to start: - 2022-blake3-* ciphers need a base64-encoded key of an exact byte length per cipher. fillProtocolDefaults was assigning a uuid-style string, which xray rejects as "bad key". Now the password is generated (or replaced if invalid) via random.Base64Bytes(n) sized to the chosen cipher. - Legacy ciphers (aes-256-gcm, chacha20-*, xchacha20-*) require a per-client method field in multi-user mode; model.Client has no Method, so settings.clients was stored without one and xray failed with "unsupported cipher method:". applyShadowsocksClientMethod now injects the top-level method into each client on add/update, and healShadowsocksClientMethods backfills it at xray-config-build time so existing inbounds heal on the next start. - xray/api.go ssCipherType switch was missing aes-256-gcm, which fell through to ss2022 path. - SSMethods dropdown now offers aes-256-gcm. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): preserve ClientRecord on inbound delete + filter Attached inbounds to multi-client protocols Replace the global orphan sweep in setRemoteTrafficLocked with a per-inbound diff cleanup: only delete a ClientRecord whose email disappeared from a snap-tracked inbound (i.e. a node-side delete). Inbounds that vanished entirely from the snap (e.g. admin deleted the inbound on master) aren't iterated, so a client whose last attachment came from that inbound is now left alone instead of being deleted alongside the inbound. ClientFormModal and ClientBulkAddModal now filter the Attached inbounds dropdown to protocols that actually support multiple clients: shadowsocks, vless, vmess, trojan, hysteria, hysteria2, and portfallback (which routes through VLESS settings). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): make empty-state text readable on dark/ultra themes The "No clients yet" empty state had a hardcoded black color (rgba(0,0,0,0.45)) that vanished against the dark backgrounds. Drop the inline color, let it inherit from the AntD theme, and fade with opacity like the mobile card empty state already does. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): client-first tgbot add flow, tgId field, lightweight inbound options - tgbot: drop legacy per-protocol Add Client UI in favour of a client-first multi-inbound flow. New BuildClientDraftMessage / getInboundsAttachPicker let an admin pick one or more inbounds and submit a single client; per- protocol secrets are now generated server-side via fillProtocolDefaults. Drops awaiting_id/awaiting_password_tr/awaiting_password_sh state cases and add_client_ch_default_id/pass_tr/pass_sh/flow callbacks. Adds a setTGUser button + awaiting_tg_id state so the bot can set Client.TgID during Add. - clients UI: add Telegram user ID input to ClientFormModal (0 = none). Hide IP Limit field entirely when ipLimitEnable is off — disabled fields still take layout space, this collapses Auth(Hysteria) to full width. - inbounds API: new GET /panel/api/inbounds/options that returns just {id, remark, protocol, port, tlsFlowCapable}. Used by the clients page pickers so the dropdown payload stays small on panels with thousands of clients (drops settings JSON, clientStats, streamSettings). Server-side TlsFlowCapable mirrors Inbound.canEnableTlsFlow so the modal no longer needs to parse streamSettings client-side. - clientInfoMsg now shows attached inbound remarks, and getInboundUsages reports the attached client count per inbound. - api-docs: document the new /options endpoint and add tgId / flow to the clients add/update bodies. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(inbounds): keep Node column visible for node-attached inbounds The Node column was bound to hasActiveNode, so disabling every node hid the column even when inbounds were still attached to those nodes — the admin lost the visual cue that those inbounds belonged to a node and would come back when it was re-enabled. Combine hasActiveNode with a new hasNodeAttachedInbound check (any dbInbound with nodeId != null) so the column survives node-disable. * fix(api-docs): accept functional-component icons in EndpointSection AntD-Vue icons (SafetyCertificateOutlined, etc.) are functional components, so the icon prop's type: Object validator was rejecting them with a "Expected Object, got Function" warning at runtime. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * test: cover crypto, random, netsafe, sub helpers, xray equals, websocket hub, node service Adds ~110 unit tests across previously untested packages. Focus on pure-logic and concurrency surfaces where regressions would silently affect users: - util/crypto, util/random: password hashing round-trip, ss2022 key generation, alphabet/length invariants. - util/netsafe: IsBlockedIP edge cases, NormalizeHost validation, SSRF guard with AllowPrivate context bypass. - util/common, util/json_util: traffic formatter, Combine nil-skip, RawMessage empty-as-null and copy-on-unmarshal. - sub: splitLinkLines, searchKey/searchHost, kcp share fields, finalmask normalization, buildVmessLink round-trip. - xray: Config.Equals and InboundConfig.Equals field-by-field, getRequiredUserString/getOptionalUserString type checks. - web/websocket: hub registration, throttling, slow-client eviction, nil-receiver safety, concurrent register/unregister. - web/service: NodeService.normalize validation, normalizeBasePath, HeartbeatPatch.ToUI mapping. - web/job: atomicBool concurrent set/takeAndReset semantics. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * i18n(clients): replace English fallbacks with proper translation keys Pulls every hard-coded English label/title in the Clients page and its four modals through the i18n layer so localized panels stop leaking English. New keys live under pages.clients (auth, hysteriaAuth, uuid, flow, flowNone, reverseTag, reverseTagPlaceholder, telegramId, telegramIdPlaceholder, created, updated, ipLimit) plus refresh at the root and toasts.bulkDeletedMixed / bulkCreatedMixed for partial-failure toasts. Also switches the add-client modal's primary button from "Add" to "Create" for consistency with other create flows. The bulk-add Random/Random+Prefix/... email-method options stay hard-coded by request - they're identifier-shaped strings. * i18n: backfill 99 missing keys across all 12 non-English locales Brings every translation file up to parity with en-US.json so the Clients page, the fallback-children inbound section, the new refresh verb, the Nodes panel-version label and a handful of older holes stop falling through to the English fallback. New strings span: - pages.clients.* (labels, confirmations, toasts, emailMethods) - pages.inbounds.portFallback.* (Reality fallback inbound section) - pages.nodes.panelVersion, menu.clients, refresh Technical identifiers (Auth, UUID, Flow, Reverse tag) are intentionally left untranslated since they correspond to xray-core field names. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * i18n: drop stale pages.client block duplicated in every non-English locale Every non-English locale carried a pages.client (singular) section with 30 entries that duplicated pages.clients (plural). The plural namespace is what the Vue code actually consumes; the singular one was dead weight from an older rename that never got cleaned up in the non-English files. Removing it brings every locale to exactly 984 keys, matching en-US.json. * chore: apply modernize analyzer fixes across codebase Mechanical replacements suggested by golang.org/x/tools/.../modernize: strings.Cut/CutPrefix/SplitSeq, slices.Contains, maps.Copy, min(), range-over-int, new(expr), strings.Builder for hot += loops, reflect.TypeFor[T](), sync.WaitGroup.Go(), drop legacy //+build lines. * feat(database): add PostgreSQL as an optional backend alongside SQLite Lets operators with large client counts or multi-node setups pick PostgreSQL at install time without breaking the existing SQLite default. Backend is selected at runtime via XUI_DB_TYPE/XUI_DB_DSN, a small dialect layer keeps the five JSON_EXTRACT/JSON_EACH queries portable, and a new `x-ui migrate-db` subcommand copies SQLite data into PostgreSQL in FK-aware order. * fix(inbounds): gate node selector to multi-node-capable protocols Hide the Deploy-To selector and clear nodeId when switching to a protocol that can't run on a remote node. Also: - subs: return 404 (not 400) when subId matches no inbounds, so VPN clients distinguish "deleted/unknown" from a server error - hysteria link gen: use the inbound's resolved address so node-managed inbounds advertise the node host instead of the central panel - shadowsocks: default network to 'tcp' (udp was causing issues for some clients on first-create) - vite dev proxy: rewrite migrated-route bypass against the live base path instead of a hardcoded single-segment regex Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): bulk add/delete correctness + perf, working pagination, delayed-start in form Bulk add/delete were serial on the frontend (one toast per call, N round-trips) and the backend race exposed by parallelizing them lost client attachments and hit UNIQUE constraint failed on client_inbounds. The single add/edit modal also had no Start-After-First-Use option, and the table never showed the delayed duration. Backend (web/service/client.go): - Per-inbound mutex on Add/Update/Del InboundClient so concurrent writers on the same inbound don't lose the read-modify-write of settings JSON. - SyncInbound skips create+join when the email is tombstoned so a concurrent maintenance pass (adjustTraffics, autoRenewClients, markClientsDisabledIn- Settings) that did a stale RMW can't resurrect a just-deleted client with a fresh id. - compactOrphans sweeps settings.clients entries whose ClientRecord no longer exists, applied in Add/DelInboundClient + DelInboundClientByEmail so each user-initiated mutation self-heals the inbound's settings. - DelInboundClient uses Pluck instead of First for the stats lookup so a missing row doesn't abort the delete with a noisy ErrRecordNotFound log. Frontend: - HttpUtil.{get,post} accept a silent option that suppresses the auto-toast. - ClientBulkAddModal fires creates in parallel + silent + one summary toast. - useClients.removeMany runs deletes in parallel + silent and refreshes once; ClientsPage bulk delete uses it and shows one aggregate toast. - useClients.applyInvalidate debounces 200 ms so the burst of N WebSocket invalidate events from the backend collapses into a single refresh. - ClientsPage pagination is reactive (paginationState ref + tablePagination computed); onTableChange persists page-size and page changes. - ClientFormModal gains a Start-After-First-Use switch + Duration days input alongside the existing Expiry Date picker; on edit-mode open a negative expiryTime is decoded back to delayed mode + days; on submit the payload sends -86400000 * days or the absolute timestamp. - ClientsPage table shows the delayed-start duration (blue tag Nd, tooltip Start After First Use: Nd) instead of infinity. - Telegram ID field in the form is hidden when /panel/setting/defaultSettings reports tgBotEnable=false; Comment then fills the row. - Form row 3 collapses UUID (span 12) + Total GB (span 8) + Limit IP (span 4) when ipLimitEnable is on, else UUID + Total GB at 12/12. - useInbounds.rollupClients counts only clients with a matching clientStats row, so orphans in settings.clients no longer inflate the inbound's count. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(windows): clean shutdown, working panel restart, harden kernel32 load Three Windows-specific issues addressed: 1. Orphaned xray-windows-amd64 after VS Code debugger stop. Delve's "Stop" sends TerminateProcess to the Go binary, which is uncatchable — our signal handlers never run, so xrayService.StopXray() is skipped and xray is left dangling. Spawn xray as a child of a Job Object with JOB_OBJECT_LIMIT_KILL_ON_JOB_CLOSE so the OS kills xray when our handle to the job is closed (which happens even on TerminateProcess). Also trap os.Interrupt in main so Ctrl+C in the terminal runs the graceful path. 2. /panel/setting/restartPanel logged "failed to send SIGHUP signal: not supported by windows" because Windows can't deliver arbitrary signals. Add a restart hook in web/global; main registers it to push SIGHUP into its own signal channel, and RestartPanel calls the hook before falling back to the (Unix-only) signal path. Same restart-loop code runs in both cases. 3. util/sys/sys_windows.go now uses windows.NewLazySystemDLL so the kernel32.dll resolve is pinned to %SystemRoot%\System32 (prevents DLL hijacking by a planted DLL next to the binary). Local filetime type replaced with windows.Filetime, and the unreliable syscall.GetLastError() fallback replaced with a type assertion on the errno captured at call time. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(sys): correct CPU/connection accounting on linux + darwin util/sys/sys_linux.go: - GetTCPCount/GetUDPCount were counting the column header row in /proc/net/{tcp,udp}[6] as a connection, inflating the reported total by 1 per non-empty file (so the panel status line always showed 2 more connections than actually existed). Replace getLinesNum + safeGetLinesNum with a single bufio.Scanner-based countConnections that skips the header. - CPUPercentRaw now opens HostProc("stat") instead of a hardcoded /proc/stat so HOST_PROC overrides apply, matching the connection counters in the same file. - Simplify CPU field unpacking: pad nums to 8 once instead of guarding every assignment with a len check. util/sys/sys_darwin.go: - Fix swapped idle/intr indices on kern.cp_time. BSD CPUSTATES order is user, nice, sys, intr, idle (CP_INTR=3, CP_IDLE=4) — gopsutil's cpu_darwin_nocgo.go reads the same layout. The previous code used out[3] as idle and out[4] as intr, so busy = total - dIdle was actually subtracting interrupt time, making the panel report CPU usage close to 100% on macOS regardless of actual load. - Collapse the per-field delta math into a single loop. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(xray): rotate crash reports into log folder, prevent overwrites writeCrashReport had two flaws: it wrote to the bin folder (alongside the xray binary) which conflates artifacts, and the second-precision timestamp meant a tight restart-loop crash burst overwrote prior reports. Write to the log folder with nanosecond precision and keep the last 10 reports. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * revert(inbounds): drop unreleased portfallback protocol The Port-with-Fallback inbound (commit 62fd9f9d) was confusing as a standalone protocol — fallbacks belong on a regular VLESS/Trojan TCP-TLS inbound, the way Xray models them natively. Rip out the entire feature cleanly (no migration needed since it was never released): protocol constant, fallback children DB table, FallbackService, 2 API endpoints, all UI rows, related translations and api-docs. A native fallback flow attached to VLESS/Trojan TCP-TLS/Reality will land in a follow-up commit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds): native fallbacks on VLESS/Trojan TCP-TLS, with working child links A VLESS or Trojan inbound on TCP with TLS or Reality can now act as a fallback master: pick existing inbounds as children and the panel auto- fills the SNI / ALPN / path / xver routing fields from each child's transport, auto-builds settings.fallbacks at config-gen time, and rewrites the child's client-share link so it advertises the master's reachable endpoint and TLS state instead of the child's loopback listen. Layout matches the Xray All-in-One Nginx example: master at :443 with clients + TLS, each child on 127.0.0.1 with its own transport+clients. Order matters (Xray walks fallbacks top-to-bottom) — reorder via the per-row up/down arrows. Path / SNI / ALPN are exposed under a per-row Edit toggle for the rare cases where the auto-derivation needs overriding; otherwise just pick a child and you're done. Backend: new InboundFallback table + FallbackService (GetByMaster / SetByMaster / GetParentForChild / BuildFallbacksJSON); two routes (GET / POST /panel/api/inbounds/:id/fallbacks); xray.GetXrayConfig injects settings.fallbacks for any VLESS/Trojan TCP-TLS/Reality inbound; GetInbounds annotates each child with FallbackParent so the frontend can rewrite links without an extra round-trip. Link projection covers every emission path — clients-page QR/links, per-inbound Get URL, raw subscription, sub-JSON, sub-Clash, and the inbounds-page link/info/QR — via a shared projectThroughFallbackMaster on the backend and a shared projectChildThroughMaster on the frontend that both handle the panel-tracked relationship and the legacy unix-socket (@vless-ws) convention. Strings translated into all 12 non-English locales. * docs: rewrite CONTRIBUTING with full local-dev setup The prior three-line CONTRIBUTING left newcomers guessing at every non-trivial step: which Go / Node versions, where xray comes from, why the panel goes blank when XUI_DEBUG=true is flipped on, how the Vue multi-page setup is wired, what to do on Windows when go build trips on the CGo SQLite driver. Now covers prerequisites, MinGW-w64 install on Windows (niXman builds or MSYS2), one-shot first-time setup, two frontend dev workflows with the XUI_DEBUG asset-cache gotcha called out, the architecture and conventions of the Vue side, a project-layout map, useful env vars, and the PR checklist. ---------
2026-05-19 03:16:42 -07:00
fi
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
if [[ "$db_choice" == "2" ]]; then
echo ""
echo -e "${green}PostgreSQL backup & restore is built into the panel:${plain}"
echo -e " ${blue}${SSL_SCHEME}://${SSL_HOST}:${config_port}/${config_webBasePath}${plain} → Backup & Restore"
echo -e "${yellow} Back Up downloads a pg_dump .dump file; Restore reloads it via pg_restore.${plain}"
fi
Random PostgreSQL role + post-install credentials display (#4608) * feat(install): random PostgreSQL role + post-install credentials display The local-Postgres installer used to bake in a static role name (`xui`) and only printed `PostgreSQL (xui@127.0.0.1:5432/xui)` at the end of install, leaving operators without the random password or any hint of how to connect from the shell. Two changes: - install_postgres_local now generates an 8-char random role name alongside the random password, and double-quotes identifiers in the CREATE/ALTER statements (a random alphanumeric may start with a digit, which Postgres rejects for unquoted identifiers). - After a successful local install, a dedicated "PostgreSQL Credentials" block is rendered in the summary — DB / user / pass / host / port / DSN / env-file path, plus ready-to-paste psql commands for both the postgres superuser and the new role. Credentials cross the subshell boundary via a 0600 tmpfile (PG_CRED_FILE) that the parent shell sources and unlinks; the PG_* vars are unset after display. Only fires for the local-install flow; the external-DSN path is unchanged. * fix(install): address Copilot review on Postgres install flow - Use mktemp (unguessable, 0600) instead of /tmp/x-ui-pg-creds.$$ and cleanup in both success and failure paths to close the symlink/race attack on the predictable filename. - In install_postgres_local, capture the prior umask and restore it after writing PG_CRED_FILE; return 1 if the write fails so the caller does not source nothing and label the install with empty PG_* vars. - On reinstall, reuse the existing xui DB owner instead of generating a fresh role each run, so existing tables stay accessible after a re-run; only the password is rotated. Falls back to a fresh random role when the DB does not exist or is owned by postgres.
2026-05-28 08:20:16 -07:00
if [[ "$db_choice" == "2" && "$pg_local_installed" == "1" ]]; then
echo ""
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${green} PostgreSQL Credentials ${plain}"
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${green}DB Name: ${PG_DB}${plain}"
echo -e "${green}Username: ${PG_USER}${plain}"
echo -e "${green}Password: ${PG_PASS}${plain}"
echo -e "${green}Host: ${PG_HOST}${plain}"
echo -e "${green}Port: ${PG_PORT}${plain}"
echo -e "${green}DSN: ${xui_dsn}${plain}"
echo -e "${green}Env file: ${xui_env_file}${plain}"
echo -e "${green}-------------------------------------------${plain}"
echo -e "${green}Connect from this server:${plain}"
echo -e " ${blue}sudo -u postgres psql -d ${PG_DB}${plain} (as the postgres superuser)"
echo -e " ${blue}PGPASSWORD='${PG_PASS}' psql -h ${PG_HOST} -p ${PG_PORT} -U ${PG_USER} -d ${PG_DB}${plain}"
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${yellow}⚠ The panel reads these credentials from ${xui_env_file}.${plain}"
echo -e "${yellow}⚠ Save the password — it is not stored anywhere else in plain text.${plain}"
unset PG_USER PG_PASS PG_HOST PG_PORT PG_DB
fi
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
# Persist a machine-parseable credentials file for cloud-init / MOTD.
: "${SSL_SCHEME:=https}"
: "${SSL_HOST:=${server_ip}}"
local db_type_out="sqlite"
[[ "$db_choice" == "2" ]] && db_type_out="postgres"
write_install_result "${config_username}" "${config_password}" "${config_port}" \
"${config_webBasePath}" "${SSL_SCHEME}" "${SSL_HOST}" "${config_apiToken}" "${db_type_out}"
Feat/multi inbound clients (#4469) * feat(clients): add shadow tables for first-class client promotion Introduces three new GORM-backed tables (clients, client_inbounds, inbound_fallback_children) and a populate-only seeder that backfills them from each inbound's existing settings.clients JSON. Duplicate emails across inbounds auto-merge under one client row, with each field conflict logged. Existing services are unchanged and continue reading from settings.clients — this commit is groundwork only. * feat(clients): make clients+client_inbounds the runtime source of truth Adds ClientService.SyncInbound that reconciles the new tables from each inbound's clients list whenever existing service paths mutate settings.clients. Wires it into AddInbound, UpdateInbound, AddInboundClient, UpdateInboundClient, DelInboundClient, DelInboundClientByEmail, DelDepletedClients, autoRenewClients, and the timestamp-backfill path in adjustTraffics, plus DetachInbound on DelInbound. GetXrayConfig now builds settings.clients from the new tables before writing config.json, and getInboundsBySubId joins through them instead of JSON_EACH on settings JSON. Live Xray config and subscription endpoints are now driven by the relational view; settings.clients JSON stays in step as a side effect of every write. * feat(clients): add top-level Clients tab and CRUD API Adds /panel/api/clients endpoints (list, get, add, update, del, attach, detach) backed by ClientService methods that orchestrate the per-inbound Add/Update/Del flows so a single client row is created once and attached to many inbounds in one operation. The frontend gains a dedicated Clients page (frontend/clients.html + src/pages/clients/) with an AntD table, multi-inbound attach modal, and full CRUD. Axios interceptor learns to honour Content-Type: application/json so the JSON endpoints work alongside the legacy form-encoded ones. The legacy per-inbound client modal stays untouched in this PR — both flows now write to the same source of truth. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds): add Port-with-Fallback inbound type Adds a new "portfallback" protocol that emits as a VLESS-TLS inbound under the hood but is paired with a sidecar table of child inbounds. Panel auto-builds settings.fallbacks at Xray-config-gen time from the sidecar — each child's listen+port becomes the fallback dest, with SNI/ALPN/path/xver match criteria pulled from the row. No more typing loopback ports by hand or keeping settings.fallbacks in sync. Backend: new FallbackService (Get/SetChildren, BuildFallbacksJSON); two new routes (GET/POST /panel/api/inbounds/:id/fallbackChildren); xray.GetXrayConfig injects fallbacks for PortFallback inbounds; the inbound model emits protocol="vless" so Xray accepts the config. Frontend: PORTFALLBACK joins the protocol dropdown; selecting it shows the standard VLESS controls plus a Fallback Children table (inbound picker + per-row SNI/ALPN/path/xver). Children are loaded on edit and replaced atomically on save. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Reset Traffic, QR Code, Info actions + Online/Remaining columns The Clients page table gains: - Online column — green/grey tag driven by /panel/api/inbounds/onlines, polled every 10s. - Remaining column — bytes-remaining tag, coloured green/orange/red against quota, purple infinity when unlimited. - Action icons per row: QR, Info, Reset traffic, Edit, Delete. ClientInfoModal shows the full client detail (uuid/password/auth, traffic ↑/↓ + remaining + all-time, expiry absolute + relative, attached inbounds chip list, online + last-online). ClientQrModal fetches links for the client's subId via /panel/api/inbounds/getSubLinks/:subId and renders each one through the existing QrPanel component. Reset Traffic confirms then calls the existing per-inbound endpoint on the client's first attached inbound (the traffic row is keyed on email globally, so any attached inbound resets the shared counter). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): expose Attached inbounds in edit mode The multi-select was gated on add-only, so editing a client had no way to change which inbounds it belonged to. The picker now shows in both modes, and on submit the modal diffs the picked set against the original attachedIds — additions go through the /attach endpoint, removals through /detach, both after the field update lands so the new attachments get the latest values. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): unbreak template parsing + stale i18n keys - InboundFormModal: split the multi-line help string in the PortFallback section onto one line — Vue's template parser was bailing on Unterminated string constant because a single-quoted literal spanned two lines inside a {{ }} interpolation. - ClientInfoModal: t('disable') was missing at the root level, so vue-i18n returned the key path literally. Use t('disabled') which exists. - Linter cleanup elsewhere: pages.client.* references renamed to pages.clients.* to match the merged i18n block; whitespace normalisation in a few unrelated Vue templates. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * 1 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(traffic): drop all-time traffic tracking Removes the AllTime field from Inbound and ClientTraffic and migrates existing DBs by dropping the all_time columns on startup. The counter duplicated up+down without adding signal, and the per-event accumulator ran on every traffic write. Frontend: drop the All-time column from the inbound list and the client-row table, the All-time row from the client info modal, and the All-Time Total Usage tile from the inbounds summary card. The allTimeTraffic/allTimeTrafficUsage i18n keys are removed across every locale. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): mobile cards, multi-select, bulk add Adds the same row-card layout the inbounds page uses on mobile: the table is suppressed under the mobile breakpoint and each client renders as a compact card with a status dot, email, Info button, Enable switch, and overflow menu. All the per-client detail (traffic, remaining, expiry, attached inbounds, flow, created/updated, URL, subscription) opens through the existing info modal. Multi-select with bulk delete wires AntD row-selection on desktop and a per-card checkbox on mobile; a Delete (N) button appears in the toolbar when anything is selected. Bulk add reuses the five email-generation modes from the inbound bulk modal but takes a multi-inbound picker so one bulk run can attach to several inbounds at once. Submits client-by-client through the existing /panel/api/clients/add endpoint. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(inbounds): remove legacy per-inbound client UI Now that clients live as first-class rows attached to one or many inbounds, the per-inbound client UI on the inbounds page is dead weight — every client action either has a global equivalent on the Clients page or makes no sense in a many-to-many world. Deletes ClientFormModal, ClientBulkModal, CopyClientsModal, and ClientRowTable from inbounds/. Strips the matching emits, refs, handlers, and dropdown menu items from InboundList and InboundsPage, and removes the dead mobile expand-chevron state and the desktop expanded-row plumbing that drove the inline client table. The InboundFormModal Clients tab still works in add-mode (one inline client at inbound creation) — that flow goes through ClientService. SyncInbound on save and remains useful. Fixes a stray "</a-dropdown>" left over by an earlier toolbar edit in ClientsPage that broke the template parser. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Delete depleted action Mirrors the legacy delDepletedClients action that lived under the inbounds page, but as a first-class /panel/api/clients/delDepleted endpoint backed by ClientService. The new path goes through ClientService.Delete for each depleted email, so the new clients + client_inbounds + xray_client_traffic tables stay consistent. Adds a danger-styled toolbar button on the Clients page (next to Reset all client traffic) with a confirm dialog and a toast reporting the deleted count. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(api): move every client-shaped endpoint off /inbounds onto /clients After the multi-inbound client migration, client state belongs to the client API surface, not the inbound one. Twelve routes that were crammed under /panel/api/inbounds/* now live where they belong, under /panel/api/clients/*. Moved (route, handler, doc): POST /clientIps/:email POST /clearClientIps/:email POST /onlines POST /lastOnline POST /updateClientTraffic/:email POST /resetAllClientTraffics/:id POST /delDepletedClients/:id POST /:id/resetClientTraffic/:email GET /getClientTraffics/:email GET /getClientTrafficsById/:id GET /getSubLinks/:subId GET /getClientLinks/:id/:email Their /clients/* counterparts are: POST /clients/clientIps/:email POST /clients/clearClientIps/:email POST /clients/onlines POST /clients/lastOnline POST /clients/updateTraffic/:email POST /clients/resetTraffic/:email (email-only, fans out) GET /clients/traffic/:email GET /clients/traffic/byId/:id GET /clients/subLinks/:subId GET /clients/links/:id/:email per-inbound resetAllClientTraffics and delDepletedClients are dropped entirely — the Clients page already exposes global Reset All Traffic and Delete depleted actions, and per-inbound resets are meaningless once a client can be attached to many inbounds. ClientService.ResetTrafficByEmail is the new email-only reset path: it looks up every inbound the client is attached to and pushes the counter reset + Xray re-add through inboundService.ResetClientTraffic for each one, so depleted users come back online instantly. Frontend callers (ClientsPage, useClients, ClientQrModal, ClientInfoModal, InboundInfoModal, InboundsPage, useInbounds) all switched to the new paths. The Inbounds page drops its per-inbound "Reset client traffic" and "Delete depleted clients" dropdown items — users do those at the client level now. api-docs is rebuilt to match. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(service): switch tgbot + ldap callers to ClientService Adds two thin helpers to ClientService (CreateOne, DetachByEmail) and rewrites tgbot.SubmitAddClient and ldap_sync_job to call ClientService directly. Removes the JSON-blob payloads (BuildJSONForProtocol output for add, clientsToJSON/clientToJSON helpers) that callers previously fed to InboundService.AddInboundClient/DelInboundClient. ldap_sync_job.batchSetEnable now loops InboundService.SetClientEnableByEmail per email instead of trying to coerce AddInboundClient into doing the update — the old path would have failed duplicate-email validation for existing clients anyway. The legacy InboundService.AddInboundClient/UpdateInboundClient/ DelInboundClient methods stay in place; they are now only used internally by ClientService Create/Update/Delete/Attach. Inlining + deleting them follows in a separate commit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(service): move all client mutation methods to ClientService Moves the client mutation surface out of InboundService and into ClientService. These methods all operate on a single client (identity fields, traffic limits, expiry, ip limit, enable state, telegram tg id) and didn't belong on the inbound aggregate. Moved (12 methods): AddInboundClient, UpdateInboundClient, DelInboundClient, DelInboundClientByEmail, checkEmailsExistForClients, SetClientTelegramUserID, checkIsEnabledByEmail, ToggleClientEnableByEmail, SetClientEnableByEmail, ResetClientIpLimitByEmail, ResetClientExpiryTimeByEmail, ResetClientTrafficLimitByEmail. Each method now takes an explicit *InboundService for the helpers that legitimately stay on InboundService (GetInbound, GetClients, runtimeFor, AddClientStat / UpdateClientStat / DelClientStat, DelClientIPs / UpdateClientIPs, emailUsedByOtherInbounds, getAllEmailSubIDs, GetClientInboundByEmail / GetClientInboundByTrafficID, GetClientTrafficByEmail). Stays on InboundService: ResetClientTrafficByEmail and ResetClientTraffic(id, email) — these mutate xray_client_traffic rows, not client identity, so they're inbound-side bookkeeping. Callers updated: tgbot (6 calls), ldap_sync_job (1 call), InboundService internal (writeBackClientSubID, CopyInboundClients, AddInbound's email-uniqueness check), ClientService Create/Update/ Delete/Attach/Detach. Also removes a dead resetAllClientTraffics controller handler whose route was already gone after the previous /clients API migration. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(clients): finish migrating to ClientService + tidy IP routes Two related cleanups in the new /clients surface: 1. Move ResetAllClientTraffics (bulk-reset of xray_client_traffic + last_traffic_reset_time, with node-runtime propagation) from InboundService to ClientService. PeriodicTrafficResetJob now holds a clientService and calls j.clientService.ResetAllClientTraffics(&j.inboundService, id). The last client-mutation method on InboundService is gone. 2. Shorten redundantly-named routes/handlers under /panel/api/clients: - /clientIps/:email -> /ips/:email (handler getIps) - /clearClientIps/:email -> /clearIps/:email (handler clearIps) The "client" prefix was redundant inside the clients namespace. Frontend (InboundInfoModal) and api-docs updated to match. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds,clients): clean up inbound modal + enrich client modal Inbound modal rework (InboundFormModal.vue + inbound.js): - Drop the embedded Client subform in the Protocol tab. Multi-inbound clients are managed exclusively from the Clients page now; a fresh inbound is created with zero clients (settings constructors default to []) and the user attaches clients afterwards. - Hide the Protocol tab entirely when it has nothing to render (VMESS, Trojan without fallbacks, Hysteria). Auto-switches active tab to Basic when the tab disappears while focused. - Move the Security section (Security selector + TLS block with certs and ECH + Reality block) out of the Stream tab into its own Security tab, sharing the canEnableStream gate. Client modal additions (ClientFormModal.vue + ClientBulkAddModal.vue): - Flow select (xtls-rprx-vision / -udp443) appears only when the panel actually has a Vision-capable inbound (VLESS or PortFallback on TCP with TLS or Reality). Hidden otherwise, and cleared when it disappears. - IP Limit input is disabled when the panel-level ipLimitEnable setting is off, fetched into useClients alongside subSettings and threaded through ClientsPage to both modals. - Edit modal now shows an "IP Log" section listing IPs that have connected with the client's credentials, with refresh and clear buttons (calls the renamed /panel/api/clients/ips and /clearIps endpoints). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(inbounds): drop manual Fallbacks UI from inbound modal The PortFallback protocol type now covers the common VLESS-master-plus-children case with auto-wired dests, so the manual Fallbacks editor (showFallbacks block in the Protocol tab) is mostly redundant. Removed: - the v-if="showFallbacks" template block (SNI/ALPN/Path/dest/PROXY rows) - the showFallbacks computed - the addFallback / delFallback helpers - the .fallbacks-header / .fallbacks-title styles - the showFallbacks gate from hasProtocolTabContent (so Trojan-over-TCP no longer shows an empty Protocol tab) Power users who need a non-inbound fallback dest (nginx, static site) can still author settings.fallbacks via the Advanced JSON tab. * feat(clients,inbounds): move search/filter to Clients page + small fixes Search/filter relocation: - Remove the search/filter toolbar (search switch + filter radio + protocol/node selects + the visibleInbounds projection + inboundsFilterState localStorage + filter CSS + the SearchOutlined/ FilterOutlined/ObjectUtil/Inbound imports it required) from InboundList. The filters were all client-oriented buckets bolted onto the inbound row. - Add a search/filter toolbar to ClientsPage with the same shape: switch between deep-text search and bucket filter (active / deactive / depleted / expiring / online) + protocol filter that matches clients attached to at least one inbound with the chosen protocol. State persists in clientsFilterState localStorage. filteredClients drives both the desktop table and the mobile card list, and select-all / allSelected / someSelected only span the visible subset. - useClients now also fetches expireDiff and trafficDiff from /panel/setting/defaultSettings (used to detect the expiring bucket); ClientsPage threads them into the client-bucket helper. Loose fixes folded in: - Add Client: email field is auto-filled with a random handle on open, matching uuid/subId/password/auth. - Inbound clone: parse and reuse the source settings JSON (with clients reset to []) instead of building a fresh defaulted Settings, so VLESS Encryption/Decryption and other non-client fields survive the clone. - en-US.json: add the ipLog string used by the edit-client modal. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Reverse tag field for VLESS-attached clients Mirrors the Flow field's pattern: a Reverse tag input appears in the Add/Edit Client modal whenever at least one selected inbound is VLESS or PortFallback. The value rides over the wire as client.reverse = { tag: '...' } so it lands directly in model.Client's *ClientReverse field; an empty value omits the reverse key entirely. On edit the field is hydrated from props.client.reverse?.tag, and the showReverseTag watcher clears the field if the user drops the last VLESS-like inbound from the selection. * fix(xray): emit only protocol-relevant fields per client entry The Xray config synthesizer was writing every identifier field (id, password, flow, auth, security/method, reverse) on every client entry regardless of the inbound's protocol. Xray ignores unknown fields, so the config worked, but it diverged from the spec and leaked secrets across protocols when one client was attached to multiple inbounds — a VLESS inbound's generated config carried the same client's Trojan password and Hysteria auth alongside its uuid. Switch on inbound.Protocol when building each entry: - VLESS / PortFallback: id, flow, reverse - VMess: id, security - Trojan: password, flow - Shadowsocks: password, method - Hysteria / Hysteria2: auth email is emitted for every protocol. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): restore auto-disable kick under new schema disableInvalidClients still resolved (inbound_tag, email) pairs via JSON_EACH(inbounds.settings.clients), which is empty after migrating to the clients + client_inbounds tables. Result: xrayApi.RemoveUser never ran for depleted clients, clients.enable stayed true so the UI showed them as active, and only xray_client_traffic.enable got flipped - making "Restart Xray After Auto Disable" only half-work. Resolve the targets via a JOIN through the new schema, flip clients.enable so the Clients page reflects the state, and drop the legacy JSON write-back plus the subId cascade workaround (email is unique now). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): live WebSocket updates + Ended status surfacing ClientsPage now subscribes to traffic / client_stats / invalidate WebSocket events instead of polling /onlines every 10s. Per-row traffic counters refresh in place, online state stays current, and list-level mutations elsewhere trigger a refresh. The client roll-up summary moves from InboundsPage to ClientsPage where it belongs, restructured into six labeled stat tiles (Total / Online / Ended / Expiring / Disabled / Active) with email popovers on the ones with issues. Auto-disabled clients (traffic exhausted or expiry passed) now classify as 'depleted' even though clients.enable=false, so they show up under the Ended filter and render a red Ended tag instead of looking indistinguishable from an operator-disabled row. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(nodes): per-node client roll-up and panel version Added transient inboundCount / clientCount / onlineCount / depletedCount fields to model.Node, populated by NodeService.GetAll via aggregated queries (one join across inbounds + client_inbounds, one over client_traffics intersected with the in-memory online emails). The Nodes list renders these as colored chips on a new "Clients" column so an operator can see at a glance how many users each node carries and how many are currently online or depleted. Also exposes the remote panel's version. The central panel adds panelVersion to its /api/server/status payload (sourced from config.GetVersion). Probe reads that field and persists it on the node row, mirroring how xrayVersion already flows. NodesPage gets a new column next to Xray Version, in both desktop and mobile views, with English and Persian strings. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): stop node sync from resurrecting deleted clients Several related issues around node-managed clients: - Remote runtime: drop the per-inbound resetAllClientTraffics path and point traffic/onlines/lastOnline fetches at the new /panel/api/clients/* routes. - Delete from master: always push the updated inbound to the node even when the client was already disabled or depleted, so the node actually loses the user instead of silently keeping it. - setRemoteTraffic: mirror remote clients into the central tables only on first discovery of a node inbound. Matched inbounds let the master own the join table, so a stale snap can no longer re-create a ClientRecord (and join row) for a client that was just deleted on the master. - ClientService.Delete: route through submitTrafficWrite so deletes serialize with node traffic merges, and switch the final ClientRecord delete to an explicit Where("id = ?") clause. - setRemoteTraffic UNIQUE-constraint fix: use clause.OnConflict on inserts and email-keyed UPDATEs for client_traffics, so mirroring a snap doesn't trip the unique email index. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(clients): switch client API endpoints from id to email All client-scoped routes now use the unique email as the path key (get, update, del, attach, detach, links). Email is the stable, protocol-independent identifier — UUIDs don't exist for trojan or shadowsocks, and internal numeric ids leaked panel implementation detail into the public API. Removed the redundant /traffic/byId/:id endpoint (covered by /traffic/:email) and collapsed /links/:id/:email into /links/:email, which now returns links across every attached inbound for the client. Frontend selection, bulk delete, and toggle state are now keyed by email as well, dropping the id→email lookup workaround. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(server): move cached state and helpers into ServerService ServerController had grown to hold its own status cache, version-list TTL cache, history-bucket whitelist, and the loop that drove all three — concerns that belong in the service layer. Pull them out: - lastStatus + the @2s refresh become ServerService.RefreshStatus and ServerService.LastStatus; the controller's cron now just orchestrates the cross-service side effects (xrayMetrics sample, websocket broadcast). - The 15-minute Xray-versions cache (with stale-on-error fallback) moves into ServerService.GetXrayVersionsCached, collapsing the controller handler to a single call. - The freedom/blackhole outbound-tag walk used by /xraylogs becomes ServerService.GetDefaultLogOutboundTags. - The allowed-history-bucket whitelist moves to package-level service.IsAllowedHistoryBucket, so both NodeController and ServerController validate against the same list. Net result: web/controller/server.go drops from 458 to 365 lines and contains only HTTP wiring + presentation-y side effects. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(api): emit JSON-text columns as nested objects Inbound, ClientRecord, and InboundClientIps store settings / streamSettings / sniffing / reverse / ips as JSON-text in the DB. The API was passing that text through verbatim, so every consumer had to JSON.parse a string inside a string. Add MarshalJSON / UnmarshalJSON so the wire format is a real nested object, while still accepting the legacy escaped-string shape on write. Frontend dbinbound.js gets a matching coerceInboundJsonField helper for the same dual-shape read path, and inbound.js toJson stops emitting empty/placeholder fields (externalProxy [], sniffing destOverride when disabled, etc.) so the new normalised JSON stays terse. api-docs and the inbound-clone path are updated to the new shape. Controller route lists are regrouped so all GETs sit above POSTs. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): include inboundIds and traffic in /clients/list ClientRecord got its own MarshalJSON in the previous commit, and ClientWithAttachments embeds it to add inboundIds and traffic. Go promotes the embedded MarshalJSON to the outer struct, so the encoder was calling ClientRecord.MarshalJSON for the whole value and silently dropping the extras. The frontend reads row.inboundIds / row.traffic from /clients/list, so attached inbounds didn't render and newly added clients looked like they hadn't saved. Add an explicit MarshalJSON on ClientWithAttachments that splices the extras in. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): gate IP Log on ipLimitEnable + clean access-log dropdown Legacy panel hid the IP Log section when access logging was off; the Vue 3 migration left it gated on isEdit only, so the section showed even when xray's access log was 'none' and nothing was being recorded. Restore the ipLimitEnable gate on the edit modal's IP Log form-item. While here, clean up the Xray Settings access-log dropdown: previously two 'none' entries appeared (an empty value labelled with t('none') and the literal 'none' from the options array). Drop the empty option for access log (the literal 'none' covers it) and relabel the empty option for error log / mask address to t('empty') so they're distinguishable. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(nodes): route per-client ops through node clients API + orphan sweep Adds Runtime methods AddClient, UpdateUser, and DeleteUser so master mutates clients on a node via /panel/api/clients/{add,update,del} rather than pushing the whole inbound. The previous rt.UpdateInbound path made the node DelInbound+AddInbound on every single-client change, briefly cycling every other user on the same inbound. DelInbound no longer filters by enable=true, so a disabled node inbound actually gets removed from the node instead of being resurrected by the next snap. setRemoteTrafficLocked now sweeps any ClientRecord with zero ClientInbound rows after SyncInbound rebuilds the attachments, which is how a node-side delete propagates back to master instead of leaving a detached ghost. ClientService.Delete tombstones the email first so a snap arriving mid-delete can't re-create the record. WebSocket broadcasts an "invalidate(clients)" message on every client mutation so the Clients page refreshes without manual reload. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(balancers): allow fallback on all strategies + feed burstObservatory from random/roundRobin Drops the random/roundRobin gate on the Fallback field in BalancerFormModal so every strategy can pick a fallback outbound. syncObservatories now feeds burstObservatory from leastLoad + random + roundRobin balancers (was leastLoad only), matching how leastPing feeds observatory. Fix the JsonEditor "Unexpected end of JSON input" that appeared when switching a balancer between leastPing and another strategy: the obsView watcher was gated on showObsEditor (a boolean OR of the two flags) and missed the case where one observatory swapped for the other in the same tick. Watch the individual flags instead so obsView flips to the surviving editor and the getter stops pointing at a deleted key. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(inbounds): use sortedInbounds for mobile empty-state check InboundList referenced an undefined visibleInbounds in the mobile card list's empty-state guard, throwing "Cannot read properties of undefined (reading 'length')" and breaking the entire mobile render. * feat(clients): sortable table columns Adds the same sortState / sortableCol / sortFns pattern InboundList uses, wrapping filteredClients in sortedClients so sort composes with the existing search/filter pipeline. Sortable: enable, email, inboundIds (attachment count), traffic, remaining, expiryTime; actions and online stay unsorted. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(shadowsocks): generate valid ss2022 keys and per-client method for legacy ciphers The Add Client flow on shadowsocks inbounds was producing xray configs that failed to start: - 2022-blake3-* ciphers need a base64-encoded key of an exact byte length per cipher. fillProtocolDefaults was assigning a uuid-style string, which xray rejects as "bad key". Now the password is generated (or replaced if invalid) via random.Base64Bytes(n) sized to the chosen cipher. - Legacy ciphers (aes-256-gcm, chacha20-*, xchacha20-*) require a per-client method field in multi-user mode; model.Client has no Method, so settings.clients was stored without one and xray failed with "unsupported cipher method:". applyShadowsocksClientMethod now injects the top-level method into each client on add/update, and healShadowsocksClientMethods backfills it at xray-config-build time so existing inbounds heal on the next start. - xray/api.go ssCipherType switch was missing aes-256-gcm, which fell through to ss2022 path. - SSMethods dropdown now offers aes-256-gcm. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): preserve ClientRecord on inbound delete + filter Attached inbounds to multi-client protocols Replace the global orphan sweep in setRemoteTrafficLocked with a per-inbound diff cleanup: only delete a ClientRecord whose email disappeared from a snap-tracked inbound (i.e. a node-side delete). Inbounds that vanished entirely from the snap (e.g. admin deleted the inbound on master) aren't iterated, so a client whose last attachment came from that inbound is now left alone instead of being deleted alongside the inbound. ClientFormModal and ClientBulkAddModal now filter the Attached inbounds dropdown to protocols that actually support multiple clients: shadowsocks, vless, vmess, trojan, hysteria, hysteria2, and portfallback (which routes through VLESS settings). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): make empty-state text readable on dark/ultra themes The "No clients yet" empty state had a hardcoded black color (rgba(0,0,0,0.45)) that vanished against the dark backgrounds. Drop the inline color, let it inherit from the AntD theme, and fade with opacity like the mobile card empty state already does. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): client-first tgbot add flow, tgId field, lightweight inbound options - tgbot: drop legacy per-protocol Add Client UI in favour of a client-first multi-inbound flow. New BuildClientDraftMessage / getInboundsAttachPicker let an admin pick one or more inbounds and submit a single client; per- protocol secrets are now generated server-side via fillProtocolDefaults. Drops awaiting_id/awaiting_password_tr/awaiting_password_sh state cases and add_client_ch_default_id/pass_tr/pass_sh/flow callbacks. Adds a setTGUser button + awaiting_tg_id state so the bot can set Client.TgID during Add. - clients UI: add Telegram user ID input to ClientFormModal (0 = none). Hide IP Limit field entirely when ipLimitEnable is off — disabled fields still take layout space, this collapses Auth(Hysteria) to full width. - inbounds API: new GET /panel/api/inbounds/options that returns just {id, remark, protocol, port, tlsFlowCapable}. Used by the clients page pickers so the dropdown payload stays small on panels with thousands of clients (drops settings JSON, clientStats, streamSettings). Server-side TlsFlowCapable mirrors Inbound.canEnableTlsFlow so the modal no longer needs to parse streamSettings client-side. - clientInfoMsg now shows attached inbound remarks, and getInboundUsages reports the attached client count per inbound. - api-docs: document the new /options endpoint and add tgId / flow to the clients add/update bodies. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(inbounds): keep Node column visible for node-attached inbounds The Node column was bound to hasActiveNode, so disabling every node hid the column even when inbounds were still attached to those nodes — the admin lost the visual cue that those inbounds belonged to a node and would come back when it was re-enabled. Combine hasActiveNode with a new hasNodeAttachedInbound check (any dbInbound with nodeId != null) so the column survives node-disable. * fix(api-docs): accept functional-component icons in EndpointSection AntD-Vue icons (SafetyCertificateOutlined, etc.) are functional components, so the icon prop's type: Object validator was rejecting them with a "Expected Object, got Function" warning at runtime. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * test: cover crypto, random, netsafe, sub helpers, xray equals, websocket hub, node service Adds ~110 unit tests across previously untested packages. Focus on pure-logic and concurrency surfaces where regressions would silently affect users: - util/crypto, util/random: password hashing round-trip, ss2022 key generation, alphabet/length invariants. - util/netsafe: IsBlockedIP edge cases, NormalizeHost validation, SSRF guard with AllowPrivate context bypass. - util/common, util/json_util: traffic formatter, Combine nil-skip, RawMessage empty-as-null and copy-on-unmarshal. - sub: splitLinkLines, searchKey/searchHost, kcp share fields, finalmask normalization, buildVmessLink round-trip. - xray: Config.Equals and InboundConfig.Equals field-by-field, getRequiredUserString/getOptionalUserString type checks. - web/websocket: hub registration, throttling, slow-client eviction, nil-receiver safety, concurrent register/unregister. - web/service: NodeService.normalize validation, normalizeBasePath, HeartbeatPatch.ToUI mapping. - web/job: atomicBool concurrent set/takeAndReset semantics. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * i18n(clients): replace English fallbacks with proper translation keys Pulls every hard-coded English label/title in the Clients page and its four modals through the i18n layer so localized panels stop leaking English. New keys live under pages.clients (auth, hysteriaAuth, uuid, flow, flowNone, reverseTag, reverseTagPlaceholder, telegramId, telegramIdPlaceholder, created, updated, ipLimit) plus refresh at the root and toasts.bulkDeletedMixed / bulkCreatedMixed for partial-failure toasts. Also switches the add-client modal's primary button from "Add" to "Create" for consistency with other create flows. The bulk-add Random/Random+Prefix/... email-method options stay hard-coded by request - they're identifier-shaped strings. * i18n: backfill 99 missing keys across all 12 non-English locales Brings every translation file up to parity with en-US.json so the Clients page, the fallback-children inbound section, the new refresh verb, the Nodes panel-version label and a handful of older holes stop falling through to the English fallback. New strings span: - pages.clients.* (labels, confirmations, toasts, emailMethods) - pages.inbounds.portFallback.* (Reality fallback inbound section) - pages.nodes.panelVersion, menu.clients, refresh Technical identifiers (Auth, UUID, Flow, Reverse tag) are intentionally left untranslated since they correspond to xray-core field names. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * i18n: drop stale pages.client block duplicated in every non-English locale Every non-English locale carried a pages.client (singular) section with 30 entries that duplicated pages.clients (plural). The plural namespace is what the Vue code actually consumes; the singular one was dead weight from an older rename that never got cleaned up in the non-English files. Removing it brings every locale to exactly 984 keys, matching en-US.json. * chore: apply modernize analyzer fixes across codebase Mechanical replacements suggested by golang.org/x/tools/.../modernize: strings.Cut/CutPrefix/SplitSeq, slices.Contains, maps.Copy, min(), range-over-int, new(expr), strings.Builder for hot += loops, reflect.TypeFor[T](), sync.WaitGroup.Go(), drop legacy //+build lines. * feat(database): add PostgreSQL as an optional backend alongside SQLite Lets operators with large client counts or multi-node setups pick PostgreSQL at install time without breaking the existing SQLite default. Backend is selected at runtime via XUI_DB_TYPE/XUI_DB_DSN, a small dialect layer keeps the five JSON_EXTRACT/JSON_EACH queries portable, and a new `x-ui migrate-db` subcommand copies SQLite data into PostgreSQL in FK-aware order. * fix(inbounds): gate node selector to multi-node-capable protocols Hide the Deploy-To selector and clear nodeId when switching to a protocol that can't run on a remote node. Also: - subs: return 404 (not 400) when subId matches no inbounds, so VPN clients distinguish "deleted/unknown" from a server error - hysteria link gen: use the inbound's resolved address so node-managed inbounds advertise the node host instead of the central panel - shadowsocks: default network to 'tcp' (udp was causing issues for some clients on first-create) - vite dev proxy: rewrite migrated-route bypass against the live base path instead of a hardcoded single-segment regex Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): bulk add/delete correctness + perf, working pagination, delayed-start in form Bulk add/delete were serial on the frontend (one toast per call, N round-trips) and the backend race exposed by parallelizing them lost client attachments and hit UNIQUE constraint failed on client_inbounds. The single add/edit modal also had no Start-After-First-Use option, and the table never showed the delayed duration. Backend (web/service/client.go): - Per-inbound mutex on Add/Update/Del InboundClient so concurrent writers on the same inbound don't lose the read-modify-write of settings JSON. - SyncInbound skips create+join when the email is tombstoned so a concurrent maintenance pass (adjustTraffics, autoRenewClients, markClientsDisabledIn- Settings) that did a stale RMW can't resurrect a just-deleted client with a fresh id. - compactOrphans sweeps settings.clients entries whose ClientRecord no longer exists, applied in Add/DelInboundClient + DelInboundClientByEmail so each user-initiated mutation self-heals the inbound's settings. - DelInboundClient uses Pluck instead of First for the stats lookup so a missing row doesn't abort the delete with a noisy ErrRecordNotFound log. Frontend: - HttpUtil.{get,post} accept a silent option that suppresses the auto-toast. - ClientBulkAddModal fires creates in parallel + silent + one summary toast. - useClients.removeMany runs deletes in parallel + silent and refreshes once; ClientsPage bulk delete uses it and shows one aggregate toast. - useClients.applyInvalidate debounces 200 ms so the burst of N WebSocket invalidate events from the backend collapses into a single refresh. - ClientsPage pagination is reactive (paginationState ref + tablePagination computed); onTableChange persists page-size and page changes. - ClientFormModal gains a Start-After-First-Use switch + Duration days input alongside the existing Expiry Date picker; on edit-mode open a negative expiryTime is decoded back to delayed mode + days; on submit the payload sends -86400000 * days or the absolute timestamp. - ClientsPage table shows the delayed-start duration (blue tag Nd, tooltip Start After First Use: Nd) instead of infinity. - Telegram ID field in the form is hidden when /panel/setting/defaultSettings reports tgBotEnable=false; Comment then fills the row. - Form row 3 collapses UUID (span 12) + Total GB (span 8) + Limit IP (span 4) when ipLimitEnable is on, else UUID + Total GB at 12/12. - useInbounds.rollupClients counts only clients with a matching clientStats row, so orphans in settings.clients no longer inflate the inbound's count. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(windows): clean shutdown, working panel restart, harden kernel32 load Three Windows-specific issues addressed: 1. Orphaned xray-windows-amd64 after VS Code debugger stop. Delve's "Stop" sends TerminateProcess to the Go binary, which is uncatchable — our signal handlers never run, so xrayService.StopXray() is skipped and xray is left dangling. Spawn xray as a child of a Job Object with JOB_OBJECT_LIMIT_KILL_ON_JOB_CLOSE so the OS kills xray when our handle to the job is closed (which happens even on TerminateProcess). Also trap os.Interrupt in main so Ctrl+C in the terminal runs the graceful path. 2. /panel/setting/restartPanel logged "failed to send SIGHUP signal: not supported by windows" because Windows can't deliver arbitrary signals. Add a restart hook in web/global; main registers it to push SIGHUP into its own signal channel, and RestartPanel calls the hook before falling back to the (Unix-only) signal path. Same restart-loop code runs in both cases. 3. util/sys/sys_windows.go now uses windows.NewLazySystemDLL so the kernel32.dll resolve is pinned to %SystemRoot%\System32 (prevents DLL hijacking by a planted DLL next to the binary). Local filetime type replaced with windows.Filetime, and the unreliable syscall.GetLastError() fallback replaced with a type assertion on the errno captured at call time. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(sys): correct CPU/connection accounting on linux + darwin util/sys/sys_linux.go: - GetTCPCount/GetUDPCount were counting the column header row in /proc/net/{tcp,udp}[6] as a connection, inflating the reported total by 1 per non-empty file (so the panel status line always showed 2 more connections than actually existed). Replace getLinesNum + safeGetLinesNum with a single bufio.Scanner-based countConnections that skips the header. - CPUPercentRaw now opens HostProc("stat") instead of a hardcoded /proc/stat so HOST_PROC overrides apply, matching the connection counters in the same file. - Simplify CPU field unpacking: pad nums to 8 once instead of guarding every assignment with a len check. util/sys/sys_darwin.go: - Fix swapped idle/intr indices on kern.cp_time. BSD CPUSTATES order is user, nice, sys, intr, idle (CP_INTR=3, CP_IDLE=4) — gopsutil's cpu_darwin_nocgo.go reads the same layout. The previous code used out[3] as idle and out[4] as intr, so busy = total - dIdle was actually subtracting interrupt time, making the panel report CPU usage close to 100% on macOS regardless of actual load. - Collapse the per-field delta math into a single loop. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(xray): rotate crash reports into log folder, prevent overwrites writeCrashReport had two flaws: it wrote to the bin folder (alongside the xray binary) which conflates artifacts, and the second-precision timestamp meant a tight restart-loop crash burst overwrote prior reports. Write to the log folder with nanosecond precision and keep the last 10 reports. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * revert(inbounds): drop unreleased portfallback protocol The Port-with-Fallback inbound (commit 62fd9f9d) was confusing as a standalone protocol — fallbacks belong on a regular VLESS/Trojan TCP-TLS inbound, the way Xray models them natively. Rip out the entire feature cleanly (no migration needed since it was never released): protocol constant, fallback children DB table, FallbackService, 2 API endpoints, all UI rows, related translations and api-docs. A native fallback flow attached to VLESS/Trojan TCP-TLS/Reality will land in a follow-up commit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds): native fallbacks on VLESS/Trojan TCP-TLS, with working child links A VLESS or Trojan inbound on TCP with TLS or Reality can now act as a fallback master: pick existing inbounds as children and the panel auto- fills the SNI / ALPN / path / xver routing fields from each child's transport, auto-builds settings.fallbacks at config-gen time, and rewrites the child's client-share link so it advertises the master's reachable endpoint and TLS state instead of the child's loopback listen. Layout matches the Xray All-in-One Nginx example: master at :443 with clients + TLS, each child on 127.0.0.1 with its own transport+clients. Order matters (Xray walks fallbacks top-to-bottom) — reorder via the per-row up/down arrows. Path / SNI / ALPN are exposed under a per-row Edit toggle for the rare cases where the auto-derivation needs overriding; otherwise just pick a child and you're done. Backend: new InboundFallback table + FallbackService (GetByMaster / SetByMaster / GetParentForChild / BuildFallbacksJSON); two routes (GET / POST /panel/api/inbounds/:id/fallbacks); xray.GetXrayConfig injects settings.fallbacks for any VLESS/Trojan TCP-TLS/Reality inbound; GetInbounds annotates each child with FallbackParent so the frontend can rewrite links without an extra round-trip. Link projection covers every emission path — clients-page QR/links, per-inbound Get URL, raw subscription, sub-JSON, sub-Clash, and the inbounds-page link/info/QR — via a shared projectThroughFallbackMaster on the backend and a shared projectChildThroughMaster on the frontend that both handle the panel-tracked relationship and the legacy unix-socket (@vless-ws) convention. Strings translated into all 12 non-English locales. * docs: rewrite CONTRIBUTING with full local-dev setup The prior three-line CONTRIBUTING left newcomers guessing at every non-trivial step: which Go / Node versions, where xray comes from, why the panel goes blank when XUI_DEBUG=true is flipped on, how the Vue multi-page setup is wired, what to do on Windows when go build trips on the CGo SQLite driver. Now covers prerequisites, MinGW-w64 install on Windows (niXman builds or MSYS2), one-shot first-time setup, two frontend dev workflows with the XUI_DEBUG asset-cache gotcha called out, the architecture and conventions of the Vue side, a project-layout map, useful env vars, and the PR checklist. ---------
2026-05-19 03:16:42 -07:00
else
2026-05-19 04:09:35 -07:00
local config_webBasePath=$(gen_random_string 18)
echo -e "${yellow}WebBasePath is missing or too short. Generating a new one...${plain}"
${xui_folder}/x-ui setting -webBasePath "${config_webBasePath}"
echo -e "${green}New WebBasePath: ${config_webBasePath}${plain}"
# If the panel is already installed but no certificate is configured, prompt for SSL now
if [[ -z "${existing_cert}" ]]; then
echo ""
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${green} SSL Certificate Setup (RECOMMENDED) ${plain}"
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${yellow}Let's Encrypt now supports both domains and IP addresses!${plain}"
echo ""
prompt_and_setup_ssl "${existing_port}" "${config_webBasePath}" "${server_ip}"
echo -e "${green}Access URL: ${SSL_SCHEME}://${SSL_HOST}:${existing_port}/${config_webBasePath}${plain}"
Feat/multi inbound clients (#4469) * feat(clients): add shadow tables for first-class client promotion Introduces three new GORM-backed tables (clients, client_inbounds, inbound_fallback_children) and a populate-only seeder that backfills them from each inbound's existing settings.clients JSON. Duplicate emails across inbounds auto-merge under one client row, with each field conflict logged. Existing services are unchanged and continue reading from settings.clients — this commit is groundwork only. * feat(clients): make clients+client_inbounds the runtime source of truth Adds ClientService.SyncInbound that reconciles the new tables from each inbound's clients list whenever existing service paths mutate settings.clients. Wires it into AddInbound, UpdateInbound, AddInboundClient, UpdateInboundClient, DelInboundClient, DelInboundClientByEmail, DelDepletedClients, autoRenewClients, and the timestamp-backfill path in adjustTraffics, plus DetachInbound on DelInbound. GetXrayConfig now builds settings.clients from the new tables before writing config.json, and getInboundsBySubId joins through them instead of JSON_EACH on settings JSON. Live Xray config and subscription endpoints are now driven by the relational view; settings.clients JSON stays in step as a side effect of every write. * feat(clients): add top-level Clients tab and CRUD API Adds /panel/api/clients endpoints (list, get, add, update, del, attach, detach) backed by ClientService methods that orchestrate the per-inbound Add/Update/Del flows so a single client row is created once and attached to many inbounds in one operation. The frontend gains a dedicated Clients page (frontend/clients.html + src/pages/clients/) with an AntD table, multi-inbound attach modal, and full CRUD. Axios interceptor learns to honour Content-Type: application/json so the JSON endpoints work alongside the legacy form-encoded ones. The legacy per-inbound client modal stays untouched in this PR — both flows now write to the same source of truth. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds): add Port-with-Fallback inbound type Adds a new "portfallback" protocol that emits as a VLESS-TLS inbound under the hood but is paired with a sidecar table of child inbounds. Panel auto-builds settings.fallbacks at Xray-config-gen time from the sidecar — each child's listen+port becomes the fallback dest, with SNI/ALPN/path/xver match criteria pulled from the row. No more typing loopback ports by hand or keeping settings.fallbacks in sync. Backend: new FallbackService (Get/SetChildren, BuildFallbacksJSON); two new routes (GET/POST /panel/api/inbounds/:id/fallbackChildren); xray.GetXrayConfig injects fallbacks for PortFallback inbounds; the inbound model emits protocol="vless" so Xray accepts the config. Frontend: PORTFALLBACK joins the protocol dropdown; selecting it shows the standard VLESS controls plus a Fallback Children table (inbound picker + per-row SNI/ALPN/path/xver). Children are loaded on edit and replaced atomically on save. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Reset Traffic, QR Code, Info actions + Online/Remaining columns The Clients page table gains: - Online column — green/grey tag driven by /panel/api/inbounds/onlines, polled every 10s. - Remaining column — bytes-remaining tag, coloured green/orange/red against quota, purple infinity when unlimited. - Action icons per row: QR, Info, Reset traffic, Edit, Delete. ClientInfoModal shows the full client detail (uuid/password/auth, traffic ↑/↓ + remaining + all-time, expiry absolute + relative, attached inbounds chip list, online + last-online). ClientQrModal fetches links for the client's subId via /panel/api/inbounds/getSubLinks/:subId and renders each one through the existing QrPanel component. Reset Traffic confirms then calls the existing per-inbound endpoint on the client's first attached inbound (the traffic row is keyed on email globally, so any attached inbound resets the shared counter). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): expose Attached inbounds in edit mode The multi-select was gated on add-only, so editing a client had no way to change which inbounds it belonged to. The picker now shows in both modes, and on submit the modal diffs the picked set against the original attachedIds — additions go through the /attach endpoint, removals through /detach, both after the field update lands so the new attachments get the latest values. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): unbreak template parsing + stale i18n keys - InboundFormModal: split the multi-line help string in the PortFallback section onto one line — Vue's template parser was bailing on Unterminated string constant because a single-quoted literal spanned two lines inside a {{ }} interpolation. - ClientInfoModal: t('disable') was missing at the root level, so vue-i18n returned the key path literally. Use t('disabled') which exists. - Linter cleanup elsewhere: pages.client.* references renamed to pages.clients.* to match the merged i18n block; whitespace normalisation in a few unrelated Vue templates. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * 1 Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(traffic): drop all-time traffic tracking Removes the AllTime field from Inbound and ClientTraffic and migrates existing DBs by dropping the all_time columns on startup. The counter duplicated up+down without adding signal, and the per-event accumulator ran on every traffic write. Frontend: drop the All-time column from the inbound list and the client-row table, the All-time row from the client info modal, and the All-Time Total Usage tile from the inbounds summary card. The allTimeTraffic/allTimeTrafficUsage i18n keys are removed across every locale. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): mobile cards, multi-select, bulk add Adds the same row-card layout the inbounds page uses on mobile: the table is suppressed under the mobile breakpoint and each client renders as a compact card with a status dot, email, Info button, Enable switch, and overflow menu. All the per-client detail (traffic, remaining, expiry, attached inbounds, flow, created/updated, URL, subscription) opens through the existing info modal. Multi-select with bulk delete wires AntD row-selection on desktop and a per-card checkbox on mobile; a Delete (N) button appears in the toolbar when anything is selected. Bulk add reuses the five email-generation modes from the inbound bulk modal but takes a multi-inbound picker so one bulk run can attach to several inbounds at once. Submits client-by-client through the existing /panel/api/clients/add endpoint. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(inbounds): remove legacy per-inbound client UI Now that clients live as first-class rows attached to one or many inbounds, the per-inbound client UI on the inbounds page is dead weight — every client action either has a global equivalent on the Clients page or makes no sense in a many-to-many world. Deletes ClientFormModal, ClientBulkModal, CopyClientsModal, and ClientRowTable from inbounds/. Strips the matching emits, refs, handlers, and dropdown menu items from InboundList and InboundsPage, and removes the dead mobile expand-chevron state and the desktop expanded-row plumbing that drove the inline client table. The InboundFormModal Clients tab still works in add-mode (one inline client at inbound creation) — that flow goes through ClientService. SyncInbound on save and remains useful. Fixes a stray "</a-dropdown>" left over by an earlier toolbar edit in ClientsPage that broke the template parser. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Delete depleted action Mirrors the legacy delDepletedClients action that lived under the inbounds page, but as a first-class /panel/api/clients/delDepleted endpoint backed by ClientService. The new path goes through ClientService.Delete for each depleted email, so the new clients + client_inbounds + xray_client_traffic tables stay consistent. Adds a danger-styled toolbar button on the Clients page (next to Reset all client traffic) with a confirm dialog and a toast reporting the deleted count. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(api): move every client-shaped endpoint off /inbounds onto /clients After the multi-inbound client migration, client state belongs to the client API surface, not the inbound one. Twelve routes that were crammed under /panel/api/inbounds/* now live where they belong, under /panel/api/clients/*. Moved (route, handler, doc): POST /clientIps/:email POST /clearClientIps/:email POST /onlines POST /lastOnline POST /updateClientTraffic/:email POST /resetAllClientTraffics/:id POST /delDepletedClients/:id POST /:id/resetClientTraffic/:email GET /getClientTraffics/:email GET /getClientTrafficsById/:id GET /getSubLinks/:subId GET /getClientLinks/:id/:email Their /clients/* counterparts are: POST /clients/clientIps/:email POST /clients/clearClientIps/:email POST /clients/onlines POST /clients/lastOnline POST /clients/updateTraffic/:email POST /clients/resetTraffic/:email (email-only, fans out) GET /clients/traffic/:email GET /clients/traffic/byId/:id GET /clients/subLinks/:subId GET /clients/links/:id/:email per-inbound resetAllClientTraffics and delDepletedClients are dropped entirely — the Clients page already exposes global Reset All Traffic and Delete depleted actions, and per-inbound resets are meaningless once a client can be attached to many inbounds. ClientService.ResetTrafficByEmail is the new email-only reset path: it looks up every inbound the client is attached to and pushes the counter reset + Xray re-add through inboundService.ResetClientTraffic for each one, so depleted users come back online instantly. Frontend callers (ClientsPage, useClients, ClientQrModal, ClientInfoModal, InboundInfoModal, InboundsPage, useInbounds) all switched to the new paths. The Inbounds page drops its per-inbound "Reset client traffic" and "Delete depleted clients" dropdown items — users do those at the client level now. api-docs is rebuilt to match. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(service): switch tgbot + ldap callers to ClientService Adds two thin helpers to ClientService (CreateOne, DetachByEmail) and rewrites tgbot.SubmitAddClient and ldap_sync_job to call ClientService directly. Removes the JSON-blob payloads (BuildJSONForProtocol output for add, clientsToJSON/clientToJSON helpers) that callers previously fed to InboundService.AddInboundClient/DelInboundClient. ldap_sync_job.batchSetEnable now loops InboundService.SetClientEnableByEmail per email instead of trying to coerce AddInboundClient into doing the update — the old path would have failed duplicate-email validation for existing clients anyway. The legacy InboundService.AddInboundClient/UpdateInboundClient/ DelInboundClient methods stay in place; they are now only used internally by ClientService Create/Update/Delete/Attach. Inlining + deleting them follows in a separate commit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(service): move all client mutation methods to ClientService Moves the client mutation surface out of InboundService and into ClientService. These methods all operate on a single client (identity fields, traffic limits, expiry, ip limit, enable state, telegram tg id) and didn't belong on the inbound aggregate. Moved (12 methods): AddInboundClient, UpdateInboundClient, DelInboundClient, DelInboundClientByEmail, checkEmailsExistForClients, SetClientTelegramUserID, checkIsEnabledByEmail, ToggleClientEnableByEmail, SetClientEnableByEmail, ResetClientIpLimitByEmail, ResetClientExpiryTimeByEmail, ResetClientTrafficLimitByEmail. Each method now takes an explicit *InboundService for the helpers that legitimately stay on InboundService (GetInbound, GetClients, runtimeFor, AddClientStat / UpdateClientStat / DelClientStat, DelClientIPs / UpdateClientIPs, emailUsedByOtherInbounds, getAllEmailSubIDs, GetClientInboundByEmail / GetClientInboundByTrafficID, GetClientTrafficByEmail). Stays on InboundService: ResetClientTrafficByEmail and ResetClientTraffic(id, email) — these mutate xray_client_traffic rows, not client identity, so they're inbound-side bookkeeping. Callers updated: tgbot (6 calls), ldap_sync_job (1 call), InboundService internal (writeBackClientSubID, CopyInboundClients, AddInbound's email-uniqueness check), ClientService Create/Update/ Delete/Attach/Detach. Also removes a dead resetAllClientTraffics controller handler whose route was already gone after the previous /clients API migration. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(clients): finish migrating to ClientService + tidy IP routes Two related cleanups in the new /clients surface: 1. Move ResetAllClientTraffics (bulk-reset of xray_client_traffic + last_traffic_reset_time, with node-runtime propagation) from InboundService to ClientService. PeriodicTrafficResetJob now holds a clientService and calls j.clientService.ResetAllClientTraffics(&j.inboundService, id). The last client-mutation method on InboundService is gone. 2. Shorten redundantly-named routes/handlers under /panel/api/clients: - /clientIps/:email -> /ips/:email (handler getIps) - /clearClientIps/:email -> /clearIps/:email (handler clearIps) The "client" prefix was redundant inside the clients namespace. Frontend (InboundInfoModal) and api-docs updated to match. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds,clients): clean up inbound modal + enrich client modal Inbound modal rework (InboundFormModal.vue + inbound.js): - Drop the embedded Client subform in the Protocol tab. Multi-inbound clients are managed exclusively from the Clients page now; a fresh inbound is created with zero clients (settings constructors default to []) and the user attaches clients afterwards. - Hide the Protocol tab entirely when it has nothing to render (VMESS, Trojan without fallbacks, Hysteria). Auto-switches active tab to Basic when the tab disappears while focused. - Move the Security section (Security selector + TLS block with certs and ECH + Reality block) out of the Stream tab into its own Security tab, sharing the canEnableStream gate. Client modal additions (ClientFormModal.vue + ClientBulkAddModal.vue): - Flow select (xtls-rprx-vision / -udp443) appears only when the panel actually has a Vision-capable inbound (VLESS or PortFallback on TCP with TLS or Reality). Hidden otherwise, and cleared when it disappears. - IP Limit input is disabled when the panel-level ipLimitEnable setting is off, fetched into useClients alongside subSettings and threaded through ClientsPage to both modals. - Edit modal now shows an "IP Log" section listing IPs that have connected with the client's credentials, with refresh and clear buttons (calls the renamed /panel/api/clients/ips and /clearIps endpoints). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(inbounds): drop manual Fallbacks UI from inbound modal The PortFallback protocol type now covers the common VLESS-master-plus-children case with auto-wired dests, so the manual Fallbacks editor (showFallbacks block in the Protocol tab) is mostly redundant. Removed: - the v-if="showFallbacks" template block (SNI/ALPN/Path/dest/PROXY rows) - the showFallbacks computed - the addFallback / delFallback helpers - the .fallbacks-header / .fallbacks-title styles - the showFallbacks gate from hasProtocolTabContent (so Trojan-over-TCP no longer shows an empty Protocol tab) Power users who need a non-inbound fallback dest (nginx, static site) can still author settings.fallbacks via the Advanced JSON tab. * feat(clients,inbounds): move search/filter to Clients page + small fixes Search/filter relocation: - Remove the search/filter toolbar (search switch + filter radio + protocol/node selects + the visibleInbounds projection + inboundsFilterState localStorage + filter CSS + the SearchOutlined/ FilterOutlined/ObjectUtil/Inbound imports it required) from InboundList. The filters were all client-oriented buckets bolted onto the inbound row. - Add a search/filter toolbar to ClientsPage with the same shape: switch between deep-text search and bucket filter (active / deactive / depleted / expiring / online) + protocol filter that matches clients attached to at least one inbound with the chosen protocol. State persists in clientsFilterState localStorage. filteredClients drives both the desktop table and the mobile card list, and select-all / allSelected / someSelected only span the visible subset. - useClients now also fetches expireDiff and trafficDiff from /panel/setting/defaultSettings (used to detect the expiring bucket); ClientsPage threads them into the client-bucket helper. Loose fixes folded in: - Add Client: email field is auto-filled with a random handle on open, matching uuid/subId/password/auth. - Inbound clone: parse and reuse the source settings JSON (with clients reset to []) instead of building a fresh defaulted Settings, so VLESS Encryption/Decryption and other non-client fields survive the clone. - en-US.json: add the ipLog string used by the edit-client modal. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): add Reverse tag field for VLESS-attached clients Mirrors the Flow field's pattern: a Reverse tag input appears in the Add/Edit Client modal whenever at least one selected inbound is VLESS or PortFallback. The value rides over the wire as client.reverse = { tag: '...' } so it lands directly in model.Client's *ClientReverse field; an empty value omits the reverse key entirely. On edit the field is hydrated from props.client.reverse?.tag, and the showReverseTag watcher clears the field if the user drops the last VLESS-like inbound from the selection. * fix(xray): emit only protocol-relevant fields per client entry The Xray config synthesizer was writing every identifier field (id, password, flow, auth, security/method, reverse) on every client entry regardless of the inbound's protocol. Xray ignores unknown fields, so the config worked, but it diverged from the spec and leaked secrets across protocols when one client was attached to multiple inbounds — a VLESS inbound's generated config carried the same client's Trojan password and Hysteria auth alongside its uuid. Switch on inbound.Protocol when building each entry: - VLESS / PortFallback: id, flow, reverse - VMess: id, security - Trojan: password, flow - Shadowsocks: password, method - Hysteria / Hysteria2: auth email is emitted for every protocol. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): restore auto-disable kick under new schema disableInvalidClients still resolved (inbound_tag, email) pairs via JSON_EACH(inbounds.settings.clients), which is empty after migrating to the clients + client_inbounds tables. Result: xrayApi.RemoveUser never ran for depleted clients, clients.enable stayed true so the UI showed them as active, and only xray_client_traffic.enable got flipped - making "Restart Xray After Auto Disable" only half-work. Resolve the targets via a JOIN through the new schema, flip clients.enable so the Clients page reflects the state, and drop the legacy JSON write-back plus the subId cascade workaround (email is unique now). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): live WebSocket updates + Ended status surfacing ClientsPage now subscribes to traffic / client_stats / invalidate WebSocket events instead of polling /onlines every 10s. Per-row traffic counters refresh in place, online state stays current, and list-level mutations elsewhere trigger a refresh. The client roll-up summary moves from InboundsPage to ClientsPage where it belongs, restructured into six labeled stat tiles (Total / Online / Ended / Expiring / Disabled / Active) with email popovers on the ones with issues. Auto-disabled clients (traffic exhausted or expiry passed) now classify as 'depleted' even though clients.enable=false, so they show up under the Ended filter and render a red Ended tag instead of looking indistinguishable from an operator-disabled row. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(nodes): per-node client roll-up and panel version Added transient inboundCount / clientCount / onlineCount / depletedCount fields to model.Node, populated by NodeService.GetAll via aggregated queries (one join across inbounds + client_inbounds, one over client_traffics intersected with the in-memory online emails). The Nodes list renders these as colored chips on a new "Clients" column so an operator can see at a glance how many users each node carries and how many are currently online or depleted. Also exposes the remote panel's version. The central panel adds panelVersion to its /api/server/status payload (sourced from config.GetVersion). Probe reads that field and persists it on the node row, mirroring how xrayVersion already flows. NodesPage gets a new column next to Xray Version, in both desktop and mobile views, with English and Persian strings. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): stop node sync from resurrecting deleted clients Several related issues around node-managed clients: - Remote runtime: drop the per-inbound resetAllClientTraffics path and point traffic/onlines/lastOnline fetches at the new /panel/api/clients/* routes. - Delete from master: always push the updated inbound to the node even when the client was already disabled or depleted, so the node actually loses the user instead of silently keeping it. - setRemoteTraffic: mirror remote clients into the central tables only on first discovery of a node inbound. Matched inbounds let the master own the join table, so a stale snap can no longer re-create a ClientRecord (and join row) for a client that was just deleted on the master. - ClientService.Delete: route through submitTrafficWrite so deletes serialize with node traffic merges, and switch the final ClientRecord delete to an explicit Where("id = ?") clause. - setRemoteTraffic UNIQUE-constraint fix: use clause.OnConflict on inserts and email-keyed UPDATEs for client_traffics, so mirroring a snap doesn't trip the unique email index. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(clients): switch client API endpoints from id to email All client-scoped routes now use the unique email as the path key (get, update, del, attach, detach, links). Email is the stable, protocol-independent identifier — UUIDs don't exist for trojan or shadowsocks, and internal numeric ids leaked panel implementation detail into the public API. Removed the redundant /traffic/byId/:id endpoint (covered by /traffic/:email) and collapsed /links/:id/:email into /links/:email, which now returns links across every attached inbound for the client. Frontend selection, bulk delete, and toggle state are now keyed by email as well, dropping the id→email lookup workaround. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(server): move cached state and helpers into ServerService ServerController had grown to hold its own status cache, version-list TTL cache, history-bucket whitelist, and the loop that drove all three — concerns that belong in the service layer. Pull them out: - lastStatus + the @2s refresh become ServerService.RefreshStatus and ServerService.LastStatus; the controller's cron now just orchestrates the cross-service side effects (xrayMetrics sample, websocket broadcast). - The 15-minute Xray-versions cache (with stale-on-error fallback) moves into ServerService.GetXrayVersionsCached, collapsing the controller handler to a single call. - The freedom/blackhole outbound-tag walk used by /xraylogs becomes ServerService.GetDefaultLogOutboundTags. - The allowed-history-bucket whitelist moves to package-level service.IsAllowedHistoryBucket, so both NodeController and ServerController validate against the same list. Net result: web/controller/server.go drops from 458 to 365 lines and contains only HTTP wiring + presentation-y side effects. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * refactor(api): emit JSON-text columns as nested objects Inbound, ClientRecord, and InboundClientIps store settings / streamSettings / sniffing / reverse / ips as JSON-text in the DB. The API was passing that text through verbatim, so every consumer had to JSON.parse a string inside a string. Add MarshalJSON / UnmarshalJSON so the wire format is a real nested object, while still accepting the legacy escaped-string shape on write. Frontend dbinbound.js gets a matching coerceInboundJsonField helper for the same dual-shape read path, and inbound.js toJson stops emitting empty/placeholder fields (externalProxy [], sniffing destOverride when disabled, etc.) so the new normalised JSON stays terse. api-docs and the inbound-clone path are updated to the new shape. Controller route lists are regrouped so all GETs sit above POSTs. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): include inboundIds and traffic in /clients/list ClientRecord got its own MarshalJSON in the previous commit, and ClientWithAttachments embeds it to add inboundIds and traffic. Go promotes the embedded MarshalJSON to the outer struct, so the encoder was calling ClientRecord.MarshalJSON for the whole value and silently dropping the extras. The frontend reads row.inboundIds / row.traffic from /clients/list, so attached inbounds didn't render and newly added clients looked like they hadn't saved. Add an explicit MarshalJSON on ClientWithAttachments that splices the extras in. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): gate IP Log on ipLimitEnable + clean access-log dropdown Legacy panel hid the IP Log section when access logging was off; the Vue 3 migration left it gated on isEdit only, so the section showed even when xray's access log was 'none' and nothing was being recorded. Restore the ipLimitEnable gate on the edit modal's IP Log form-item. While here, clean up the Xray Settings access-log dropdown: previously two 'none' entries appeared (an empty value labelled with t('none') and the literal 'none' from the options array). Drop the empty option for access log (the literal 'none' covers it) and relabel the empty option for error log / mask address to t('empty') so they're distinguishable. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(nodes): route per-client ops through node clients API + orphan sweep Adds Runtime methods AddClient, UpdateUser, and DeleteUser so master mutates clients on a node via /panel/api/clients/{add,update,del} rather than pushing the whole inbound. The previous rt.UpdateInbound path made the node DelInbound+AddInbound on every single-client change, briefly cycling every other user on the same inbound. DelInbound no longer filters by enable=true, so a disabled node inbound actually gets removed from the node instead of being resurrected by the next snap. setRemoteTrafficLocked now sweeps any ClientRecord with zero ClientInbound rows after SyncInbound rebuilds the attachments, which is how a node-side delete propagates back to master instead of leaving a detached ghost. ClientService.Delete tombstones the email first so a snap arriving mid-delete can't re-create the record. WebSocket broadcasts an "invalidate(clients)" message on every client mutation so the Clients page refreshes without manual reload. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(balancers): allow fallback on all strategies + feed burstObservatory from random/roundRobin Drops the random/roundRobin gate on the Fallback field in BalancerFormModal so every strategy can pick a fallback outbound. syncObservatories now feeds burstObservatory from leastLoad + random + roundRobin balancers (was leastLoad only), matching how leastPing feeds observatory. Fix the JsonEditor "Unexpected end of JSON input" that appeared when switching a balancer between leastPing and another strategy: the obsView watcher was gated on showObsEditor (a boolean OR of the two flags) and missed the case where one observatory swapped for the other in the same tick. Watch the individual flags instead so obsView flips to the surviving editor and the getter stops pointing at a deleted key. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(inbounds): use sortedInbounds for mobile empty-state check InboundList referenced an undefined visibleInbounds in the mobile card list's empty-state guard, throwing "Cannot read properties of undefined (reading 'length')" and breaking the entire mobile render. * feat(clients): sortable table columns Adds the same sortState / sortableCol / sortFns pattern InboundList uses, wrapping filteredClients in sortedClients so sort composes with the existing search/filter pipeline. Sortable: enable, email, inboundIds (attachment count), traffic, remaining, expiryTime; actions and online stay unsorted. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(shadowsocks): generate valid ss2022 keys and per-client method for legacy ciphers The Add Client flow on shadowsocks inbounds was producing xray configs that failed to start: - 2022-blake3-* ciphers need a base64-encoded key of an exact byte length per cipher. fillProtocolDefaults was assigning a uuid-style string, which xray rejects as "bad key". Now the password is generated (or replaced if invalid) via random.Base64Bytes(n) sized to the chosen cipher. - Legacy ciphers (aes-256-gcm, chacha20-*, xchacha20-*) require a per-client method field in multi-user mode; model.Client has no Method, so settings.clients was stored without one and xray failed with "unsupported cipher method:". applyShadowsocksClientMethod now injects the top-level method into each client on add/update, and healShadowsocksClientMethods backfills it at xray-config-build time so existing inbounds heal on the next start. - xray/api.go ssCipherType switch was missing aes-256-gcm, which fell through to ss2022 path. - SSMethods dropdown now offers aes-256-gcm. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): preserve ClientRecord on inbound delete + filter Attached inbounds to multi-client protocols Replace the global orphan sweep in setRemoteTrafficLocked with a per-inbound diff cleanup: only delete a ClientRecord whose email disappeared from a snap-tracked inbound (i.e. a node-side delete). Inbounds that vanished entirely from the snap (e.g. admin deleted the inbound on master) aren't iterated, so a client whose last attachment came from that inbound is now left alone instead of being deleted alongside the inbound. ClientFormModal and ClientBulkAddModal now filter the Attached inbounds dropdown to protocols that actually support multiple clients: shadowsocks, vless, vmess, trojan, hysteria, hysteria2, and portfallback (which routes through VLESS settings). Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): make empty-state text readable on dark/ultra themes The "No clients yet" empty state had a hardcoded black color (rgba(0,0,0,0.45)) that vanished against the dark backgrounds. Drop the inline color, let it inherit from the AntD theme, and fade with opacity like the mobile card empty state already does. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(clients): client-first tgbot add flow, tgId field, lightweight inbound options - tgbot: drop legacy per-protocol Add Client UI in favour of a client-first multi-inbound flow. New BuildClientDraftMessage / getInboundsAttachPicker let an admin pick one or more inbounds and submit a single client; per- protocol secrets are now generated server-side via fillProtocolDefaults. Drops awaiting_id/awaiting_password_tr/awaiting_password_sh state cases and add_client_ch_default_id/pass_tr/pass_sh/flow callbacks. Adds a setTGUser button + awaiting_tg_id state so the bot can set Client.TgID during Add. - clients UI: add Telegram user ID input to ClientFormModal (0 = none). Hide IP Limit field entirely when ipLimitEnable is off — disabled fields still take layout space, this collapses Auth(Hysteria) to full width. - inbounds API: new GET /panel/api/inbounds/options that returns just {id, remark, protocol, port, tlsFlowCapable}. Used by the clients page pickers so the dropdown payload stays small on panels with thousands of clients (drops settings JSON, clientStats, streamSettings). Server-side TlsFlowCapable mirrors Inbound.canEnableTlsFlow so the modal no longer needs to parse streamSettings client-side. - clientInfoMsg now shows attached inbound remarks, and getInboundUsages reports the attached client count per inbound. - api-docs: document the new /options endpoint and add tgId / flow to the clients add/update bodies. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(inbounds): keep Node column visible for node-attached inbounds The Node column was bound to hasActiveNode, so disabling every node hid the column even when inbounds were still attached to those nodes — the admin lost the visual cue that those inbounds belonged to a node and would come back when it was re-enabled. Combine hasActiveNode with a new hasNodeAttachedInbound check (any dbInbound with nodeId != null) so the column survives node-disable. * fix(api-docs): accept functional-component icons in EndpointSection AntD-Vue icons (SafetyCertificateOutlined, etc.) are functional components, so the icon prop's type: Object validator was rejecting them with a "Expected Object, got Function" warning at runtime. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * test: cover crypto, random, netsafe, sub helpers, xray equals, websocket hub, node service Adds ~110 unit tests across previously untested packages. Focus on pure-logic and concurrency surfaces where regressions would silently affect users: - util/crypto, util/random: password hashing round-trip, ss2022 key generation, alphabet/length invariants. - util/netsafe: IsBlockedIP edge cases, NormalizeHost validation, SSRF guard with AllowPrivate context bypass. - util/common, util/json_util: traffic formatter, Combine nil-skip, RawMessage empty-as-null and copy-on-unmarshal. - sub: splitLinkLines, searchKey/searchHost, kcp share fields, finalmask normalization, buildVmessLink round-trip. - xray: Config.Equals and InboundConfig.Equals field-by-field, getRequiredUserString/getOptionalUserString type checks. - web/websocket: hub registration, throttling, slow-client eviction, nil-receiver safety, concurrent register/unregister. - web/service: NodeService.normalize validation, normalizeBasePath, HeartbeatPatch.ToUI mapping. - web/job: atomicBool concurrent set/takeAndReset semantics. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * i18n(clients): replace English fallbacks with proper translation keys Pulls every hard-coded English label/title in the Clients page and its four modals through the i18n layer so localized panels stop leaking English. New keys live under pages.clients (auth, hysteriaAuth, uuid, flow, flowNone, reverseTag, reverseTagPlaceholder, telegramId, telegramIdPlaceholder, created, updated, ipLimit) plus refresh at the root and toasts.bulkDeletedMixed / bulkCreatedMixed for partial-failure toasts. Also switches the add-client modal's primary button from "Add" to "Create" for consistency with other create flows. The bulk-add Random/Random+Prefix/... email-method options stay hard-coded by request - they're identifier-shaped strings. * i18n: backfill 99 missing keys across all 12 non-English locales Brings every translation file up to parity with en-US.json so the Clients page, the fallback-children inbound section, the new refresh verb, the Nodes panel-version label and a handful of older holes stop falling through to the English fallback. New strings span: - pages.clients.* (labels, confirmations, toasts, emailMethods) - pages.inbounds.portFallback.* (Reality fallback inbound section) - pages.nodes.panelVersion, menu.clients, refresh Technical identifiers (Auth, UUID, Flow, Reverse tag) are intentionally left untranslated since they correspond to xray-core field names. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * i18n: drop stale pages.client block duplicated in every non-English locale Every non-English locale carried a pages.client (singular) section with 30 entries that duplicated pages.clients (plural). The plural namespace is what the Vue code actually consumes; the singular one was dead weight from an older rename that never got cleaned up in the non-English files. Removing it brings every locale to exactly 984 keys, matching en-US.json. * chore: apply modernize analyzer fixes across codebase Mechanical replacements suggested by golang.org/x/tools/.../modernize: strings.Cut/CutPrefix/SplitSeq, slices.Contains, maps.Copy, min(), range-over-int, new(expr), strings.Builder for hot += loops, reflect.TypeFor[T](), sync.WaitGroup.Go(), drop legacy //+build lines. * feat(database): add PostgreSQL as an optional backend alongside SQLite Lets operators with large client counts or multi-node setups pick PostgreSQL at install time without breaking the existing SQLite default. Backend is selected at runtime via XUI_DB_TYPE/XUI_DB_DSN, a small dialect layer keeps the five JSON_EXTRACT/JSON_EACH queries portable, and a new `x-ui migrate-db` subcommand copies SQLite data into PostgreSQL in FK-aware order. * fix(inbounds): gate node selector to multi-node-capable protocols Hide the Deploy-To selector and clear nodeId when switching to a protocol that can't run on a remote node. Also: - subs: return 404 (not 400) when subId matches no inbounds, so VPN clients distinguish "deleted/unknown" from a server error - hysteria link gen: use the inbound's resolved address so node-managed inbounds advertise the node host instead of the central panel - shadowsocks: default network to 'tcp' (udp was causing issues for some clients on first-create) - vite dev proxy: rewrite migrated-route bypass against the live base path instead of a hardcoded single-segment regex Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(clients): bulk add/delete correctness + perf, working pagination, delayed-start in form Bulk add/delete were serial on the frontend (one toast per call, N round-trips) and the backend race exposed by parallelizing them lost client attachments and hit UNIQUE constraint failed on client_inbounds. The single add/edit modal also had no Start-After-First-Use option, and the table never showed the delayed duration. Backend (web/service/client.go): - Per-inbound mutex on Add/Update/Del InboundClient so concurrent writers on the same inbound don't lose the read-modify-write of settings JSON. - SyncInbound skips create+join when the email is tombstoned so a concurrent maintenance pass (adjustTraffics, autoRenewClients, markClientsDisabledIn- Settings) that did a stale RMW can't resurrect a just-deleted client with a fresh id. - compactOrphans sweeps settings.clients entries whose ClientRecord no longer exists, applied in Add/DelInboundClient + DelInboundClientByEmail so each user-initiated mutation self-heals the inbound's settings. - DelInboundClient uses Pluck instead of First for the stats lookup so a missing row doesn't abort the delete with a noisy ErrRecordNotFound log. Frontend: - HttpUtil.{get,post} accept a silent option that suppresses the auto-toast. - ClientBulkAddModal fires creates in parallel + silent + one summary toast. - useClients.removeMany runs deletes in parallel + silent and refreshes once; ClientsPage bulk delete uses it and shows one aggregate toast. - useClients.applyInvalidate debounces 200 ms so the burst of N WebSocket invalidate events from the backend collapses into a single refresh. - ClientsPage pagination is reactive (paginationState ref + tablePagination computed); onTableChange persists page-size and page changes. - ClientFormModal gains a Start-After-First-Use switch + Duration days input alongside the existing Expiry Date picker; on edit-mode open a negative expiryTime is decoded back to delayed mode + days; on submit the payload sends -86400000 * days or the absolute timestamp. - ClientsPage table shows the delayed-start duration (blue tag Nd, tooltip Start After First Use: Nd) instead of infinity. - Telegram ID field in the form is hidden when /panel/setting/defaultSettings reports tgBotEnable=false; Comment then fills the row. - Form row 3 collapses UUID (span 12) + Total GB (span 8) + Limit IP (span 4) when ipLimitEnable is on, else UUID + Total GB at 12/12. - useInbounds.rollupClients counts only clients with a matching clientStats row, so orphans in settings.clients no longer inflate the inbound's count. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(windows): clean shutdown, working panel restart, harden kernel32 load Three Windows-specific issues addressed: 1. Orphaned xray-windows-amd64 after VS Code debugger stop. Delve's "Stop" sends TerminateProcess to the Go binary, which is uncatchable — our signal handlers never run, so xrayService.StopXray() is skipped and xray is left dangling. Spawn xray as a child of a Job Object with JOB_OBJECT_LIMIT_KILL_ON_JOB_CLOSE so the OS kills xray when our handle to the job is closed (which happens even on TerminateProcess). Also trap os.Interrupt in main so Ctrl+C in the terminal runs the graceful path. 2. /panel/setting/restartPanel logged "failed to send SIGHUP signal: not supported by windows" because Windows can't deliver arbitrary signals. Add a restart hook in web/global; main registers it to push SIGHUP into its own signal channel, and RestartPanel calls the hook before falling back to the (Unix-only) signal path. Same restart-loop code runs in both cases. 3. util/sys/sys_windows.go now uses windows.NewLazySystemDLL so the kernel32.dll resolve is pinned to %SystemRoot%\System32 (prevents DLL hijacking by a planted DLL next to the binary). Local filetime type replaced with windows.Filetime, and the unreliable syscall.GetLastError() fallback replaced with a type assertion on the errno captured at call time. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(sys): correct CPU/connection accounting on linux + darwin util/sys/sys_linux.go: - GetTCPCount/GetUDPCount were counting the column header row in /proc/net/{tcp,udp}[6] as a connection, inflating the reported total by 1 per non-empty file (so the panel status line always showed 2 more connections than actually existed). Replace getLinesNum + safeGetLinesNum with a single bufio.Scanner-based countConnections that skips the header. - CPUPercentRaw now opens HostProc("stat") instead of a hardcoded /proc/stat so HOST_PROC overrides apply, matching the connection counters in the same file. - Simplify CPU field unpacking: pad nums to 8 once instead of guarding every assignment with a len check. util/sys/sys_darwin.go: - Fix swapped idle/intr indices on kern.cp_time. BSD CPUSTATES order is user, nice, sys, intr, idle (CP_INTR=3, CP_IDLE=4) — gopsutil's cpu_darwin_nocgo.go reads the same layout. The previous code used out[3] as idle and out[4] as intr, so busy = total - dIdle was actually subtracting interrupt time, making the panel report CPU usage close to 100% on macOS regardless of actual load. - Collapse the per-field delta math into a single loop. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * fix(xray): rotate crash reports into log folder, prevent overwrites writeCrashReport had two flaws: it wrote to the bin folder (alongside the xray binary) which conflates artifacts, and the second-precision timestamp meant a tight restart-loop crash burst overwrote prior reports. Write to the log folder with nanosecond precision and keep the last 10 reports. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * revert(inbounds): drop unreleased portfallback protocol The Port-with-Fallback inbound (commit 62fd9f9d) was confusing as a standalone protocol — fallbacks belong on a regular VLESS/Trojan TCP-TLS inbound, the way Xray models them natively. Rip out the entire feature cleanly (no migration needed since it was never released): protocol constant, fallback children DB table, FallbackService, 2 API endpoints, all UI rows, related translations and api-docs. A native fallback flow attached to VLESS/Trojan TCP-TLS/Reality will land in a follow-up commit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com> * feat(inbounds): native fallbacks on VLESS/Trojan TCP-TLS, with working child links A VLESS or Trojan inbound on TCP with TLS or Reality can now act as a fallback master: pick existing inbounds as children and the panel auto- fills the SNI / ALPN / path / xver routing fields from each child's transport, auto-builds settings.fallbacks at config-gen time, and rewrites the child's client-share link so it advertises the master's reachable endpoint and TLS state instead of the child's loopback listen. Layout matches the Xray All-in-One Nginx example: master at :443 with clients + TLS, each child on 127.0.0.1 with its own transport+clients. Order matters (Xray walks fallbacks top-to-bottom) — reorder via the per-row up/down arrows. Path / SNI / ALPN are exposed under a per-row Edit toggle for the rare cases where the auto-derivation needs overriding; otherwise just pick a child and you're done. Backend: new InboundFallback table + FallbackService (GetByMaster / SetByMaster / GetParentForChild / BuildFallbacksJSON); two routes (GET / POST /panel/api/inbounds/:id/fallbacks); xray.GetXrayConfig injects settings.fallbacks for any VLESS/Trojan TCP-TLS/Reality inbound; GetInbounds annotates each child with FallbackParent so the frontend can rewrite links without an extra round-trip. Link projection covers every emission path — clients-page QR/links, per-inbound Get URL, raw subscription, sub-JSON, sub-Clash, and the inbounds-page link/info/QR — via a shared projectThroughFallbackMaster on the backend and a shared projectChildThroughMaster on the frontend that both handle the panel-tracked relationship and the legacy unix-socket (@vless-ws) convention. Strings translated into all 12 non-English locales. * docs: rewrite CONTRIBUTING with full local-dev setup The prior three-line CONTRIBUTING left newcomers guessing at every non-trivial step: which Go / Node versions, where xray comes from, why the panel goes blank when XUI_DEBUG=true is flipped on, how the Vue multi-page setup is wired, what to do on Windows when go build trips on the CGo SQLite driver. Now covers prerequisites, MinGW-w64 install on Windows (niXman builds or MSYS2), one-shot first-time setup, two frontend dev workflows with the XUI_DEBUG asset-cache gotcha called out, the architecture and conventions of the Vue side, a project-layout map, useful env vars, and the PR checklist. ---------
2026-05-19 03:16:42 -07:00
else
2026-05-19 04:09:35 -07:00
# If a cert already exists, just show the access URL
echo -e "${green}Access URL: https://${server_ip}:${existing_port}/${config_webBasePath}${plain}"
2025-09-25 05:34:12 -07:00
fi
2023-02-09 11:18:06 -08:00
fi
else
2026-05-19 04:09:35 -07:00
if [[ "$existing_hasDefaultCredential" == "true" ]]; then
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
local config_username="${XUI_USERNAME:-$(gen_random_string 10)}"
local config_password="${XUI_PASSWORD:-$(gen_random_string 10)}"
2026-05-19 04:09:35 -07:00
echo -e "${yellow}Default credentials detected. Security update required...${plain}"
${xui_folder}/x-ui setting -username "${config_username}" -password "${config_password}"
echo -e "Generated new random login credentials:"
echo -e "###############################################"
echo -e "${green}Username: ${config_username}${plain}"
echo -e "${green}Password: ${config_password}${plain}"
echo -e "###############################################"
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
# Persist a machine-parseable credentials file for cloud-init / MOTD.
local config_apiToken
fix(cli): let -getApiToken name the token it regenerates (#6405) * fix(cli): let -getApiToken name the token it regenerates The flag's help text said "Display current API token". It cannot display anything -- tokens are stored as SHA-256 hashes, and the command's own first two output lines say so. What it does is destroy and reissue a credential: GetApiToken calls RecreateByName on the hardcoded name "cli-fallback". The help therefore invited an operator to run a command they believed was read-only, and it revoked a token someone else was holding. Because that name is a single global slot, two callers silently invalidate each other, and the loser is left with a token that answers HTTP 404 with an empty body -- indistinguishable from a wrong base path, so the failure does not even say what happened. install.sh is one of those callers, at lines 1231 and 1325, so the collision already exists inside this repository. Add -tokenName, defaulting to cli-fallback so install.sh and every existing invocation behave exactly as before. -getApiToken stays a boolean on purpose: install.sh calls it as `x-ui setting -getApiToken true`, and a string flag would swallow that trailing argument and mint a token named "true". The name now reaches both branches of GetApiToken. On a database with no tokens the command used to create one called "install", which the CLI could then never rotate -- defeating the stated purpose of the cli-fallback constant, that -getApiToken cannot accumulate admin-equivalent credentials it never revokes. Both branches use the resolved name, so repeated calls rotate a single slot instead of leaving a permanent token behind. Also cap the name at 64 characters in RecreateByName. Create already enforces that limit on the same column; RecreateByName did not, and it now receives operator input. Assisted-by: Claude Code:claude-opus-5 (mostly) * fix(cli): keep the installer's token out of the rotated slot Folding both branches of GetApiToken onto one name made the bug worse in the exact case this change is about. install.sh records the token it gets on a fresh panel; with both branches on cli-fallback, the next bare -getApiToken rotated that very row and silently invalidated the credential written into the install-result file. Restore the split default -- "install" when the database has no tokens, cli-fallback when it does -- so nothing about an unnamed call changes. An explicit -tokenName still applies to both branches, which is what keeps the flag coherent: -tokenName ci-bot now yields ci-bot on a fresh panel too, rather than "install". Pin it with a test that reads the install row's id and hash before and after a rotation, since a name-only assertion would pass against a deleted-and- recreated row. * fix(cli): stop the `-getApiToken true` form from swallowing -tokenName Three corrections from review. Go's flag package stops parsing at the first non-flag argument, so the trailing `true` in install.sh's invocation does not merely get ignored -- it terminates parsing. An operator copying that documented shape and writing `x-ui setting -getApiToken true -tokenName ci-bot` left tokenName empty, so the command rotated cli-fallback: the shared-slot collision this change exists to remove, reachable through the one form the repository itself demonstrates. Verified against the built binary, which printed `The API token "cli-fallback" has been regenerated`. Drop the stray `true` from both install.sh call sites so the documented form no longer teaches the trap, and warn whenever `setting` is given positional arguments, naming what was ignored. A warning rather than an error, because an older install.sh in the wild still passes `true` and must keep working. Cover both branches in the help strings. They described only the rotation path, so on a fresh panel -h announced cli-fallback while the command actually mints `install`, and nothing is regenerated or invalidated there at all -- misleading help being the defect this change set out to remove. Assert the concrete error in the name-length test. It checked only that some error came back, which RecreateByName's empty-name guard and its transaction errors would satisfy just as well.
2026-09-08 05:15:37 -07:00
config_apiToken=$(${xui_folder}/x-ui setting -getApiToken | grep -Eo 'apiToken: .+' | awk '{print $2}')
feat: release-driven golden-image & unattended-install deployment pipeline (#5323) * feat(install): add non-interactive install path for cloud/golden-image use Trigger non-interactive mode when XUI_NONINTERACTIVE=1 or stdin is not a TTY (curl | bash, cloud-init). Every prompt is then replaced by an env var or a sane default; interactive prompts stay byte-for-byte identical. Honored env vars: XUI_USERNAME, XUI_PASSWORD, XUI_PANEL_PORT, XUI_WEB_BASE_PATH (unset => random, as before), XUI_SSL_MODE=none|ip|domain (default none), XUI_DOMAIN, XUI_ACME_EMAIL, XUI_DB_TYPE/XUI_DB_DSN, plus additive XUI_ACME_HTTP_PORT, XUI_SSL_IPV6, XUI_SERVER_IP. On success, write /etc/x-ui/install-result.env (mode 600) with the panel creds + access URL + api token, in both interactive and non-interactive modes, so cloud-init/MOTD can surface them. Postgres in non-interactive mode requires XUI_DB_DSN or installs locally; never silently downgrades. * feat(deploy): add first-boot per-instance credential generation Golden images ship with no x-ui.db. x-ui-firstboot.sh runs once (guarded by /etc/x-ui/.firstboot-done), before x-ui.service, and replaces the seeded admin/admin with fresh random username/password on a random high port, regenerates the session secret/panel GUID via 'x-ui setting -reset', mints an API token, and writes the creds to /etc/x-ui/credentials.txt (600) + /etc/motd. Idempotent: skips regeneration if a non-default admin already exists. The oneshot unit is ordered After=network-online/cloud-init and Before=x-ui.service so the panel never serves default credentials. * chore(deploy): force LF for cloud-image deploy assets (.service/.hcl/.yaml) * feat(deploy): add Packer config + provisioning scripts for golden image One build, two sources: amazon-ebs (AWS AMI, Canonical Ubuntu 24.04 base via source_ami_filter) and qemu (qcow2 + raw, NoCloud-seeded for build-time SSH). Provisioner order is fixed: provision.sh -> harden.sh -> cleanup.sh. - provision.sh: downloads the released x-ui tarball (no Go build), installs the panel + firstboot unit, enables but does NOT start services, creates NO DB. - harden.sh: key-only SSH, no root password login, locks default account passwords, enables unattended-upgrades (scanner-compliant). - cleanup.sh: wipes any DB/creds, SSH host keys, authorized_keys, machine-id, cloud-init state, logs and history; fails the build if any secret survives. packer fmt -check clean; packer validate passes for both sources. * feat(deploy): add generic cloud-init user-data for unattended install cloud-init.yaml installs the latest 3x-ui non-interactively (XUI_NONINTERACTIVE=1) on any cloud-init platform, generating unique per-instance credentials and surfacing them via /etc/x-ui/install-result.env, serial console and MOTD. README documents per-provider usage (Hetzner/AWS/DO/Vultr/GCP/Azure/Oracle) and all XUI_* knobs. * ci: add image.yml to build cloud images on release On release: published (or workflow_dispatch with a tag), waits for the x-ui-linux-amd64.tar.gz asset (handles the release-matrix upload race), then: - qemu-image (always): builds the qcow2 with Packer and attaches a compressed .qcow2.xz + sha256 to the GitHub release. Uses KVM when /dev/kvm exists, else TCG. - ami-image (gated): builds the AWS AMI only when AWS creds exist (OIDC role preferred, else access keys), so forks skip cleanly. Prints the AMI ID to the job summary. No secrets or AMI IDs are committed. * test(deploy): add container smoke tests for install + firstboot smoke-noninteractive.sh: runs install.sh piped (no TTY) with XUI_NONINTERACTIVE=1 in an Ubuntu container; asserts install-result.env (600) holds random non-default creds, hasDefaultCredential is false, and the panel serves HTTP. smoke-firstboot.sh: installs the released binary with no DB, runs x-ui-firstboot.sh; asserts per-instance creds + credentials.txt (600) + MOTD, no admin/admin, and that a second run is a no-op (sentinel honored). smoke.yml runs both as gated jobs on PRs/pushes touching install.sh or deploy/**. Both pass locally against the v3.3.1 release binary. * docs(deploy): add Packer/marketplace docs and link from README - deploy/README.md: index of the cloud-deploy tooling and the two models - deploy/packer/README.md: how to build locally, variables, first-boot behavior - deploy/marketplace/aws/README.md: seller registration -> AMI scan -> limited-visibility preview -> go-public checklist - deploy/marketplace/hetzner/README.md: cloud-init-first guidance + snapshot caveat (delete x-ui.db first) + hetznercloud/apps reference - README.md: link the unattended-install / cloud-image docs from Quick Start * feat(deploy): build golden images for arm64 as well as amd64 The install path was already multi-arch (install.sh auto-detects arch); this extends the golden image + CI to arm64: - packer: xui_arch (amd64|arm64, validated) now derives the base AMI filter and the Ubuntu cloud image; the qemu source switches to qemu-system-aarch64 + virt machine + AAVMF UEFI firmware for arm64. amd64 path unchanged. - image.yml: arch matrix. AMIs for amd64 (t3.small) + arm64 (t4g.small/Graviton) from one runner; qcow2 for amd64 on a standard runner and arm64 on a native ubuntu-24.04-arm runner. Waits for both release tarballs. - smoke.yml: run install + firstboot smoke tests on amd64 and arm64 runners; smoke-firstboot.sh now resolves the arch tarball via dpkg. - docs updated for both arches. packer fmt/validate pass for amd64 and arm64; actionlint + shellcheck clean. Verified locally: non-interactive install AND firstboot run on the real arm64 release binary under emulation (ELF aarch64, no admin/admin). * chore(deploy): default AWS region to eu-central-1 (Frankfurt) Replace the us-east-1 fallback in image.yml (4 sites) and the Packer 'region' default + doc examples. Still overridable via the AWS_REGION repo variable / the -var 'region=...' flag. * feat(deploy): add Amazon Lightsail support (launch script + snapshot builder) Lightsail can't launch from an EC2 AMI and its blueprint list isn't self-publishable, so add the two self-service paths instead: - launch-script.sh: paste into Lightsail 'Add launch script' (or --user-data) to install 3x-ui non-interactively with unique per-instance credentials. - snapshot-userdata.sh + build-snapshot.sh: AWS CLI pipeline that provisions a build instance (panel installed, NO DB, firstboot enabled), runs the shared cleanup.sh, then snapshots it. Instances launched from the snapshot mint their own credentials on first boot. Optional --panel-port pins a known port for the Lightsail firewall. - README documents both paths, the firewall caveat, and the blueprint reality. EC2 AMI / Marketplace path kept untouched alongside. All scripts shellcheck-clean. * fix(deploy): address Copilot PR review findings - install.sh + firstboot: write install-result.env / credentials.txt values with printf %q so the files stay safe to source even if creds are pinned with shell metacharacters (no-op for the alphanumeric random defaults). - firstboot: fail closed if 'x-ui setting -show' can't be parsed to true/false — exit without writing the sentinel so the next boot retries, instead of silently skipping regeneration and risking admin/admin. - firstboot + cloud-init + lightsail launch-script: keep secrets out of the world-readable /etc/motd (show URL + username only; full creds via the mode-600 file / serial console). - lightsail build-snapshot: handle download-default-key-pair returning either a PEM or base64, and assert a valid PEM before using it for SSH. - image.yml: pin hashicorp/setup-packer@v3 (was @main). - deploy/README: document XUI_ACME_HTTP_PORT / XUI_SSL_IPV6 / XUI_SERVER_IP. Both container smoke tests still pass; shellcheck + actionlint clean.
2026-06-14 09:08:35 -07:00
: "${SSL_SCHEME:=https}"
: "${SSL_HOST:=${server_ip}}"
write_install_result "${config_username}" "${config_password}" "${existing_port}" \
"${existing_webBasePath}" "${SSL_SCHEME}" "${SSL_HOST}" "${config_apiToken}" "${XUI_DB_TYPE:-sqlite}"
2026-05-19 04:09:35 -07:00
else
echo -e "${green}Username, Password, and WebBasePath are properly set.${plain}"
fi
# Existing install: if no cert configured, prompt user for SSL setup
# Properly detect empty cert by checking if cert: line exists and has content after it
existing_cert=$(${xui_folder}/x-ui setting -getCert true | grep 'cert:' | awk -F': ' '{print $2}' | tr -d '[:space:]')
if [[ -z "$existing_cert" ]]; then
echo ""
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${green} SSL Certificate Setup (RECOMMENDED) ${plain}"
echo -e "${green}═══════════════════════════════════════════${plain}"
echo -e "${yellow}Let's Encrypt now supports both domains and IP addresses!${plain}"
echo ""
prompt_and_setup_ssl "${existing_port}" "${existing_webBasePath}" "${server_ip}"
echo -e "${green}Access URL: ${SSL_SCHEME}://${SSL_HOST}:${existing_port}/${existing_webBasePath}${plain}"
else
echo -e "${green}SSL certificate already configured. No action needed.${plain}"
fi
2025-09-25 05:34:12 -07:00
fi
2026-05-04 04:20:24 -07:00
2026-05-19 04:09:35 -07:00
${xui_folder}/x-ui migrate
}
# setup_fail2ban auto-installs and configures fail2ban for the IP Limit feature
# by invoking the freshly installed x-ui CLI. IP Limit is load-bearing on
# fail2ban (without it the panel disables the limitIp field and zeroes existing
# limits), so a fresh install should make it work out of the box, just like the
# Docker image already does. Non-fatal by design: a fail2ban failure must never
# abort the panel install.
setup_fail2ban() {
if [[ -n "${XUI_ENABLE_FAIL2BAN+x}" && "${XUI_ENABLE_FAIL2BAN}" != "true" ]]; then
echo -e "${yellow}XUI_ENABLE_FAIL2BAN=${XUI_ENABLE_FAIL2BAN}, skipping Fail2ban auto-setup.${plain}"
return 0
fi
if [[ ! -x /usr/bin/x-ui ]]; then
echo -e "${yellow}x-ui CLI not found; skipping Fail2ban auto-setup.${plain}"
return 0
fi
fix(install): fetch x-ui.sh and unit files from the installed release tag (#6391) * fix(install): fetch x-ui.sh and unit files from the installed release tag install.sh and update.sh pin the panel archive to a release tag but always took x-ui.sh, x-ui.rc and the service units from main, so the management script and the binary of one installation came from different commits: the fail2ban templates and setting flags the script writes drift silently against an older binary, two installs of the same tag differ, and a reviewed or digest-pinned installer still runs unreviewed code from main. Use the same ref as the archive, keeping main only for the rolling dev-latest build. The menu's "update menu" and update_shell paths now fetch the script matching the installed version and fall back to main with a visible notice when no script is published for it. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): fall back to main for files a pinned tag does not publish Review follow-ups. install.sh accepts tags down to v2.3.5, but x-ui.rc only exists from v2.8.4 and the split x-ui.service.* files are newer still, so pinning those to the tag made an Alpine install of an old tag 404 after the previous install was already removed. Probe the tag for each file and fall back to main with a notice when it is missing, as the menu already does for x-ui.sh. The fail2ban auto-setup probe also trusted the exit status of 'x-ui setup-fail2ban', but scripts before v3.4.0 have no such subcommand and exit 0 from the usage banner, so the installer reported a setup that never ran. Skip with a notice when the installed script does not know the subcommand. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): refuse a tag that does not publish a needed script Falling back to main reintroduced the binary/script mismatch the tag pinning exists to remove, and it fired at points where install.sh and update.sh have already stopped and removed the previous installation -- so the quiet path was also the one that could not be undone. Probe the tag instead, before anything is touched, for every file that is always fetched from GitHub (x-ui.sh, plus x-ui.rc on Alpine), and abort with the HTTP status when one is missing. The unit files stay unprobed: they are only fetched when the release tarball omits them, so an old tag that ships x-ui.service inside its tarball still installs. Their existing failure message now names the ref it tried. Also tighten the setup-fail2ban probe to the dispatcher's case arm rather than any mention of the string, which also matches a comment. --------- Co-authored-by: Claude <noreply@anthropic.com>
2026-09-03 07:50:53 -07:00
# Scripts older than v3.4.0 have no setup-fail2ban and exit 0 from the
# usage banner, which would read as success here.
if ! grep -q '"setup-fail2ban")' /usr/bin/x-ui; then
echo -e "${yellow}This x-ui.sh predates 'x-ui setup-fail2ban'; skipping Fail2ban auto-setup.${plain}"
return 0
fi
echo -e "${green}Setting up Fail2ban for the IP Limit feature...${plain}"
if /usr/bin/x-ui setup-fail2ban; then
echo -e "${green}Fail2ban setup complete.${plain}"
else
echo -e "${yellow}Fail2ban setup did not finish; IP Limit stays disabled until you run 'x-ui' and open the IP Limit menu. Continuing.${plain}"
fi
return 0
}
fix(systemd): harden shipped x-ui unit files (#6718) * fix(systemd): harden shipped x-ui unit files The units ran the panel as root with no sandboxing: systemd-analyze security rates them 9.6 UNSAFE. Add NoNewPrivileges, ProtectSystem=full with ReadWritePaths for the default XUI_DB_FOLDER/XUI_BIN_FOLDER/XUI_LOG_FOLDER stores, kernel and clock protections, UMask=0077, RestrictAddressFamilies, a CapabilityBoundingSet with NET_ADMIN/NET_BIND_SERVICE/NET_RAW and SystemCallFilter=@system-service. PrivateTmp is deliberately omitted: the web updater hands a path inside the system temp directory to a systemd-run transient unit, which does not share the service's private tmpfs. ProtectHome stays read-only because installs keep TLS certificates under the root home directory. Fixes #6605 * fix(systemd): ship ReadWriteDirectories= alias for systemd < 231 ReadWritePaths= only exists since systemd 231; install.sh still supports CentOS 7 (systemd 219), where the directive is ignored and ProtectSystem=full would leave the panel state directory read-only, breaking its database. * fix(systemd): keep root's DAC bits and regenerate the write paths Two follow-ups to the hardening, both reported by review on #6718. CAP_DAC_OVERRIDE and CAP_DAC_READ_SEARCH were dropped from the bounding set. Root holds them normally, and a bounding set is subtracted from root too: the panel could no longer read a private key it does not own (a Caddy-issued certificate under its own state dir, an acme.sh home, any 0600 file owned by another account). That fails quietly for TLS -- the panel listener logs the tls.LoadX509KeyPair error and keeps serving plain HTTP, and Xray inbounds using that key stop -- so both bits stay. ProtectSystem=full plus a hard-coded ReadWritePaths list broke installs whose XUI_DB_FOLDER/XUI_LOG_FOLDER/XUI_BIN_FOLDER live outside the defaults, and the workaround of editing the unit did not survive an update, because install.sh and update.sh reinstall the unit from the release tarball. The folders actually in use are now resolved from the same env file the unit passes to the panel and regenerated into x-ui.service.d/10-xui-write-paths.conf on every install and update, so a relocated store stays writable and the list is not reset. The unit keeps the plain-install defaults plus XUI_SERVICE, which the in-panel updater needs when systemd-run is unavailable and it falls back to a child process that inherits this sandbox while update.sh lands the unit again. Uninstall removes the drop-in with the unit. * fix(systemd): keep the seccomp whitelist off old systemd, tighten the rest Review of the previous head found that SystemCallFilter=@system-service plus SystemCallErrorNumber=EPERM is a hard regression on the platforms this PR means to keep working. @-named filter groups exist from systemd 239 on, and older systemd does not ignore an unknown group name: on <231 the name fails to resolve and the filter stays the built-in whitelist of execve/exit/exit_group/ rt_sigreturn/sigreturn, on 231..238 it degrades to @default. Either way the panel then gets EPERM on read/openat/mmap/clone and cannot start -- a CentOS 7 or Ubuntu 18.04 install would come up dead after this update. The two directives now live in the generated drop-in and are written only when "systemctl --version" reports 239 or newer, so old hosts keep the rest of the hardening and simply go without seccomp. The same review listed three more items, all addressed here: - a comment claiming ProtectSystem=full "keeps everything outside /var, /run and the listed ReadWritePaths read-only" -- that is `strict`; `full` locks down /usr, /boot, /efi and /etc; - /etc/systemd/system was granted writable for the in-panel updater's fallback, but that fallback cannot work under this sandbox at all: update.sh also stages the release archive beside the main folder, replaces /usr/bin/x-ui and calls the package manager. The entry is gone and update.sh now stops up front with one clear message when the directories it needs are read-only, instead of failing halfway with "Failed to download x-ui"; - CAP_DAC_READ_SEARCH is redundant next to CAP_DAC_OVERRIDE, so the bounding set keeps just the latter. Relocating a store by editing the env file alone is documented in the unit and in the generated drop-in: the drop-in is only written by install/update, so one of those has to be re-run afterwards. Verified with a local harness (9 checks: plain defaults, relocated store read from the env file, the same list produced by update.sh, duplicate collapse, seccomp present at systemd 249 and absent at 238, read-only guard) and bash -n on install.sh, update.sh, x-ui.sh. systemd-analyze is not available here, so the unit files themselves are unverified by a parser. * fix(systemd): actually wire the read-only guard, drop the superseded drop-in Re-review of the previous head caught two leftovers from that commit: - require_writable_update_paths was defined but never called, so the guard the unit comments, the commit message and the PR comment promise did not exist at all. It is now called at the top level, before install_base, i.e. before anything with a side effect: a sandboxed fallback run stops with one clear message instead of failing halfway, which on a relocated main folder meant the old install removed and the service folder rewritten before dying on /usr/bin. - the drop-in this branch replaced (10-xui-write-paths.conf) is no longer written or referenced, but nothing removed it either. Whoever installed the build that wrote it keeps its wider list, including the writable service folder, until it is deleted by hand. Both generators now remove it. Harness extended to 11 checks: the superseded file is gone after a run, the guard is actually called, plus the previous nine (defaults, env-file relocation, same list from update.sh, dedupe, seccomp at 249 / absent at 238, read-only guard) and bash -n on the three scripts. * fix(systemd): correct two comments and keep spaces out of the path list Second-opinion review of the previous head (two models, both asked to state platforms and versions) produced three actionable items: a wrong comment kept from the earlier commits, a wrong generalisation about the filter groups, and a path-list case that would leave the panel unable to start. - the ProtectSystem= comment claimed strict leaves /var and /run writable. It does not: strict mounts the whole hierarchy read-only and only the kernel API filesystems stay as they are. The sentence was already wrong before this branch and moving it to ProtectSystem=full did not fix it. - "the @-named filter groups need systemd >= 239" is the wrong generalisation: named groups exist since 231, it is @system-service that arrived in 239. The unit files, both script comments and the drop-in body now name the group. - a folder containing whitespace (XUI_DB_FOLDER="/srv/panel data") was written into ReadWritePaths= verbatim. That directive is a whitespace-separated list, so the entry splits into "-/srv/panel" and "data", and systemd rejects the whole drop-in: the panel then does not start at all. Such folders are left out and reported to the operator instead; the other paths are still written. Harness extended with three checks for the whitespace case (folder left out, remaining paths intact, warning emitted) and the duplicate-store case now reads its own env file instead of the previous one, so it tests what it claims. 14 checks plus bash -n on the three scripts, all passing. * fix(systemd): act on the independent review of the drop-in generator A read-only review of the branch head (another model, given the diff and the sources, asked to cite only verified lines) confirmed the earlier work and turned up four items that are fixed here: - a folder name carrying a literal % went into ReadWritePaths= as it was, and systemd expands %-specifiers in unit files: with XUI_DB_FOLDER=/srv/x%-ui the entry no longer named the directory the panel writes to and the panel could not write its database. The path is now emitted as %%; the duplicate check keeps comparing the unescaped value. - systemd older than 229/242/244 does not know NoNewPrivileges, ProtectClock, ProtectHostname and ProtectKernelLogs. It logs them and carries on, so CentOS 7 (systemd 219, which install.sh explicitly supports) runs with less hardening than the unit lists. install.sh and update.sh now print which protections need a newer systemd, which ones still apply, and that upgrading systemd is what changes it. - the updater's writability guard asked [[ -w ]] about the parent directories. It creates and removes a probe file instead, so an immutable attribute or a full filesystem is caught as well (a read-only mount was already caught). - the generator's comment claimed to resolve the folders the service actually uses, while the shipped unit hard-codes WorkingDirectory= and ExecStart= under /usr/local/x-ui. The comment now states what XUI_MAIN_FOLDER really feeds -- the location install.sh/update.sh install into and the base for a relative XUI_BIN_FOLDER -- and that a relocated main folder needs the unit edited too. Rejected from the same review, with the evidence: that [[ -w ]] cannot see a read-only mount (access(W_OK)/faccessat consults __mnt_is_readonly before the mode bits), and that the /etc ReadWritePaths entry is an exception granted for /etc rather than a default store already in the list. Harness extended: 19 checks (escaped %, the old-systemd note, whitespace and duplicate folders, seccomp gating, the read-only guard) plus bash -n on the three scripts, all passing. * fix(systemd): name the hardening old systemd really ignores The old-systemd note fired only below 239 and listed wrong versions: RHEL 8 (239) and Debian 10 (241) silently lose ProtectHostname and RestrictSUIDSGID (242), ProtectKernelLogs (244) and ProtectClock (245) with no note, while CentOS 7 was told NoNewPrivileges (187) and ProtectHome=read-only (214) were not applied although both are. The note is now built from a directive/version table taken from systemd.exec(5) and lists only what the running systemd lacks. Also drop the removal of 10-xui-write-paths.conf: only an intermediate commit of this branch wrote that file, no release ever shipped it. --------- Co-authored-by: Кот <kot@zeroclaw.local> Co-authored-by: Sanaei <ho3ein.sanaei@gmail.com>
2026-10-05 09:47:10 -07:00
# Major version of the local systemd, 0 when it cannot be determined. The
# SystemCallFilter=@system-service group only exists from systemd 239 on (other
# @-named groups exist since 231); on older versions an unknown group is not
# ignored safely, the filter stays in force and leaves a whitelist the panel
# cannot run under.
_xui_systemd_major_version() {
local version=""
if command -v systemctl > /dev/null 2>&1; then
version="$(systemctl --version 2>/dev/null | awk 'NR == 1 {print $2}')"
fi
if [[ ! "$version" =~ ^[0-9]+$ ]]; then
echo 0
return 0
fi
echo "$version"
}
# The shipped units list hardening that older systemd does not know: the
# directive is logged and ignored at load time rather than rejected, so the
# panel still starts, only without that protection. Each entry is the systemd
# release that introduced the directive (systemd.exec(5)); everything else in
# the unit predates the oldest systemd install.sh supports (CentOS 7 has 219).
# SystemCallFilter= is listed because the drop-in only writes it from 239 on.
_xui_warn_unsupported_hardening() {
local version entry missing=""
version="$(_xui_systemd_major_version)"
[[ "$version" -gt 0 ]] || return 0
for entry in RestrictRealtime:231 ReadWritePaths:231 ProtectKernelTunables:232 \
ProtectKernelModules:232 RestrictNamespaces:233 LockPersonality:235 \
SystemCallFilter:239 ProtectHostname:242 RestrictSUIDSGID:242 \
ProtectKernelLogs:244 ProtectClock:245; do
if [[ "$version" -lt "${entry##*:}" ]]; then
missing="${missing:+$missing, }${entry%%:*} (${entry##*:})"
fi
done
[[ -n "$missing" ]] || return 0
echo -e "${yellow}Note: systemd ${version} ignores part of the hardening in x-ui.service; the panel still starts.${plain}"
echo " Not applied, needs a newer systemd: ${missing}."
if [[ "$version" -lt 231 ]]; then
echo " The panel's folders stay writable through ReadWriteDirectories=, the alias this script installs."
fi
echo " The rest of the hardening is in force. Upgrade systemd to apply the above."
return 0
}
# ProtectSystem=full makes /usr, /boot, /efi and /etc read-only. ProtectSystem=
# strict would make the whole hierarchy read-only (only the kernel API
# filesystems stay as they are), and that would break the panel's own use of
# /tmp. The panel's stores are configurable (XUI_DB_FOLDER, XUI_LOG_FOLDER,
# XUI_BIN_FOLDER), and XUI_MAIN_FOLDER is the folder install.sh/update.sh place
# the files in -- the unit's WorkingDirectory on a stock install, and what a
# relative XUI_BIN_FOLDER is resolved against. So a hard-coded list in the unit
# either misses a relocated store -- the panel then cannot write its own SQLite
# database and sits in a Restart=on-failure loop -- or forces the operator to
# edit a file that every install/update overwrites from the release tarball.
# install.sh and update.sh therefore regenerate the drop-in from the folders
# actually in use, and the unit's own ReadWritePaths only carry the
# plain-install defaults. A relocated store means re-running install or update:
# the drop-in is only written here.
_xui_service_write_paths_dropin() {
# $1 is the env file to resolve the XUI_* folders from; callers pass nothing
# and get the OS-specific path the unit itself uses.
local env_file="${1:-}"
local dropin_dir dropin temp_file
local db_folder log_folder bin_folder main_folder
local path line="" whitespace_paths="" seen_paths="" escaped_path
if [[ -z "$env_file" ]]; then
case "${release}" in
ubuntu | debian | armbian)
env_file="/etc/default/x-ui"
;;
arch | manjaro | parch | alpine)
env_file="/etc/conf.d/x-ui"
;;
*)
env_file="/etc/sysconfig/x-ui"
;;
esac
fi
if [[ -r "$env_file" ]]; then
set -a
# shellcheck disable=SC1090
source "$env_file"
set +a
fi
# XUI_* wins over the script's own default: the unit hands that same env
# file to the panel through EnvironmentFile=, so these are the folders it
# will actually use.
main_folder="${XUI_MAIN_FOLDER:-${xui_folder}}"
db_folder="${XUI_DB_FOLDER:-/etc/x-ui}"
log_folder="${XUI_LOG_FOLDER:-/var/log/x-ui}"
# An empty XUI_BIN_FOLDER resolves to "bin" relative to the panel's working
# directory, which the unit sets to the main folder.
bin_folder="${XUI_BIN_FOLDER:-bin}"
if [[ "$bin_folder" != /* ]]; then
bin_folder="${main_folder%/}/${bin_folder#./}"
fi
for path in "$db_folder" "$log_folder" "$bin_folder" "$main_folder"; do
[[ "$path" == /* ]] || continue
# ReadWritePaths= is a whitespace-separated list, and a folder whose
# name contains whitespace cannot be written into it without relying on
# quoting. A wrong entry makes systemd reject the whole drop-in and the
# panel would not start, so leave such a folder out and say so instead.
if [[ "$path" != "${path//[[:space:]]/}" ]]; then
whitespace_paths="${whitespace_paths:+$whitespace_paths }$path"
continue
fi
case " $seen_paths " in
*" $path "*) continue ;;
esac
seen_paths="${seen_paths}${seen_paths:+ }$path"
# systemd expands %-specifiers in unit files, so a folder name carrying
# a literal % has to be written as %%, or the entry stops naming the
# folder systemd is meant to keep writable.
escaped_path="${path//%/%%}"
line="${line} -${escaped_path}"
done
if [[ -n "$whitespace_paths" ]]; then
echo "Warning: these folders contain whitespace and were left out of" >&2
echo " 10-xui-sandbox.conf: $whitespace_paths" >&2
echo " The panel cannot write to them under the unit's sandbox." >&2
fi
line="${line# }"
[[ -n "$line" ]] || return 1
dropin_dir="${xui_service}/x-ui.service.d"
dropin="${dropin_dir}/10-xui-sandbox.conf"
temp_file="${dropin}.tmp.$$"
mkdir -p "$dropin_dir" || return 1
cat > "$temp_file" << EOF
# Regenerated by install.sh/update.sh on every install and update: edits here
# are lost, and the list only reflects the XUI_* variables read from
# ${env_file} at that moment. Re-run install/update after moving a store.
# It lists the folders the panel writes to. Put local additions in their own
# drop-in, for example 20-x-ui-local.conf, which nothing here touches.
[Service]
ReadWritePaths=${line}
ReadWriteDirectories=${line}
EOF
if [[ "$(_xui_systemd_major_version)" -ge 239 ]]; then
cat >> "$temp_file" << 'EOF'
# @system-service needs systemd >= 239; on older versions the unknown group
# would leave the panel with a filter it cannot start under (x-ui.service.*).
SystemCallFilter=@system-service
SystemCallErrorNumber=EPERM
EOF
fi
if [[ ! -s "$temp_file" ]]; then
rm -f "$temp_file"
return 1
fi
chmod 644 "$temp_file"
mv -f "$temp_file" "$dropin" || { rm -f "$temp_file"; return 1; }
if command -v systemctl > /dev/null 2>&1; then
systemctl daemon-reload > /dev/null 2>&1 || true
fi
return 0
}
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
# Lands a systemd unit file at ${xui_service}/x-ui.service via a temp file +
# atomic mv, so a failed cp/curl or an interrupted mv never leaves a
# truncated unit file at the live path -- systemd would then fail to parse
# it on the next daemon-reload/start. Same pattern already used for
# /usr/bin/x-ui elsewhere in this script. source_is_url picks cp (from a
# file already extracted from the release tarball) vs curl (GitHub fallback).
_install_xui_service_unit() {
local source="$1"
local source_is_url="$2"
local dest="${xui_service}/x-ui.service"
local temp_file="${dest}.tmp.$$"
rm -f "$temp_file"
if [[ "$source_is_url" == "true" ]]; then
curl -fLRo "$temp_file" "$source" > /dev/null 2>&1
else
cp -f "$source" "$temp_file" > /dev/null 2>&1
fi
if [[ $? -ne 0 ]]; then
rm -f "$temp_file"
return 1
fi
if [[ ! -s "$temp_file" ]]; then
rm -f "$temp_file"
return 1
fi
mv -f "$temp_file" "$dest"
if [[ $? -ne 0 ]]; then
rm -f "$temp_file"
return 1
fi
fix(systemd): harden shipped x-ui unit files (#6718) * fix(systemd): harden shipped x-ui unit files The units ran the panel as root with no sandboxing: systemd-analyze security rates them 9.6 UNSAFE. Add NoNewPrivileges, ProtectSystem=full with ReadWritePaths for the default XUI_DB_FOLDER/XUI_BIN_FOLDER/XUI_LOG_FOLDER stores, kernel and clock protections, UMask=0077, RestrictAddressFamilies, a CapabilityBoundingSet with NET_ADMIN/NET_BIND_SERVICE/NET_RAW and SystemCallFilter=@system-service. PrivateTmp is deliberately omitted: the web updater hands a path inside the system temp directory to a systemd-run transient unit, which does not share the service's private tmpfs. ProtectHome stays read-only because installs keep TLS certificates under the root home directory. Fixes #6605 * fix(systemd): ship ReadWriteDirectories= alias for systemd < 231 ReadWritePaths= only exists since systemd 231; install.sh still supports CentOS 7 (systemd 219), where the directive is ignored and ProtectSystem=full would leave the panel state directory read-only, breaking its database. * fix(systemd): keep root's DAC bits and regenerate the write paths Two follow-ups to the hardening, both reported by review on #6718. CAP_DAC_OVERRIDE and CAP_DAC_READ_SEARCH were dropped from the bounding set. Root holds them normally, and a bounding set is subtracted from root too: the panel could no longer read a private key it does not own (a Caddy-issued certificate under its own state dir, an acme.sh home, any 0600 file owned by another account). That fails quietly for TLS -- the panel listener logs the tls.LoadX509KeyPair error and keeps serving plain HTTP, and Xray inbounds using that key stop -- so both bits stay. ProtectSystem=full plus a hard-coded ReadWritePaths list broke installs whose XUI_DB_FOLDER/XUI_LOG_FOLDER/XUI_BIN_FOLDER live outside the defaults, and the workaround of editing the unit did not survive an update, because install.sh and update.sh reinstall the unit from the release tarball. The folders actually in use are now resolved from the same env file the unit passes to the panel and regenerated into x-ui.service.d/10-xui-write-paths.conf on every install and update, so a relocated store stays writable and the list is not reset. The unit keeps the plain-install defaults plus XUI_SERVICE, which the in-panel updater needs when systemd-run is unavailable and it falls back to a child process that inherits this sandbox while update.sh lands the unit again. Uninstall removes the drop-in with the unit. * fix(systemd): keep the seccomp whitelist off old systemd, tighten the rest Review of the previous head found that SystemCallFilter=@system-service plus SystemCallErrorNumber=EPERM is a hard regression on the platforms this PR means to keep working. @-named filter groups exist from systemd 239 on, and older systemd does not ignore an unknown group name: on <231 the name fails to resolve and the filter stays the built-in whitelist of execve/exit/exit_group/ rt_sigreturn/sigreturn, on 231..238 it degrades to @default. Either way the panel then gets EPERM on read/openat/mmap/clone and cannot start -- a CentOS 7 or Ubuntu 18.04 install would come up dead after this update. The two directives now live in the generated drop-in and are written only when "systemctl --version" reports 239 or newer, so old hosts keep the rest of the hardening and simply go without seccomp. The same review listed three more items, all addressed here: - a comment claiming ProtectSystem=full "keeps everything outside /var, /run and the listed ReadWritePaths read-only" -- that is `strict`; `full` locks down /usr, /boot, /efi and /etc; - /etc/systemd/system was granted writable for the in-panel updater's fallback, but that fallback cannot work under this sandbox at all: update.sh also stages the release archive beside the main folder, replaces /usr/bin/x-ui and calls the package manager. The entry is gone and update.sh now stops up front with one clear message when the directories it needs are read-only, instead of failing halfway with "Failed to download x-ui"; - CAP_DAC_READ_SEARCH is redundant next to CAP_DAC_OVERRIDE, so the bounding set keeps just the latter. Relocating a store by editing the env file alone is documented in the unit and in the generated drop-in: the drop-in is only written by install/update, so one of those has to be re-run afterwards. Verified with a local harness (9 checks: plain defaults, relocated store read from the env file, the same list produced by update.sh, duplicate collapse, seccomp present at systemd 249 and absent at 238, read-only guard) and bash -n on install.sh, update.sh, x-ui.sh. systemd-analyze is not available here, so the unit files themselves are unverified by a parser. * fix(systemd): actually wire the read-only guard, drop the superseded drop-in Re-review of the previous head caught two leftovers from that commit: - require_writable_update_paths was defined but never called, so the guard the unit comments, the commit message and the PR comment promise did not exist at all. It is now called at the top level, before install_base, i.e. before anything with a side effect: a sandboxed fallback run stops with one clear message instead of failing halfway, which on a relocated main folder meant the old install removed and the service folder rewritten before dying on /usr/bin. - the drop-in this branch replaced (10-xui-write-paths.conf) is no longer written or referenced, but nothing removed it either. Whoever installed the build that wrote it keeps its wider list, including the writable service folder, until it is deleted by hand. Both generators now remove it. Harness extended to 11 checks: the superseded file is gone after a run, the guard is actually called, plus the previous nine (defaults, env-file relocation, same list from update.sh, dedupe, seccomp at 249 / absent at 238, read-only guard) and bash -n on the three scripts. * fix(systemd): correct two comments and keep spaces out of the path list Second-opinion review of the previous head (two models, both asked to state platforms and versions) produced three actionable items: a wrong comment kept from the earlier commits, a wrong generalisation about the filter groups, and a path-list case that would leave the panel unable to start. - the ProtectSystem= comment claimed strict leaves /var and /run writable. It does not: strict mounts the whole hierarchy read-only and only the kernel API filesystems stay as they are. The sentence was already wrong before this branch and moving it to ProtectSystem=full did not fix it. - "the @-named filter groups need systemd >= 239" is the wrong generalisation: named groups exist since 231, it is @system-service that arrived in 239. The unit files, both script comments and the drop-in body now name the group. - a folder containing whitespace (XUI_DB_FOLDER="/srv/panel data") was written into ReadWritePaths= verbatim. That directive is a whitespace-separated list, so the entry splits into "-/srv/panel" and "data", and systemd rejects the whole drop-in: the panel then does not start at all. Such folders are left out and reported to the operator instead; the other paths are still written. Harness extended with three checks for the whitespace case (folder left out, remaining paths intact, warning emitted) and the duplicate-store case now reads its own env file instead of the previous one, so it tests what it claims. 14 checks plus bash -n on the three scripts, all passing. * fix(systemd): act on the independent review of the drop-in generator A read-only review of the branch head (another model, given the diff and the sources, asked to cite only verified lines) confirmed the earlier work and turned up four items that are fixed here: - a folder name carrying a literal % went into ReadWritePaths= as it was, and systemd expands %-specifiers in unit files: with XUI_DB_FOLDER=/srv/x%-ui the entry no longer named the directory the panel writes to and the panel could not write its database. The path is now emitted as %%; the duplicate check keeps comparing the unescaped value. - systemd older than 229/242/244 does not know NoNewPrivileges, ProtectClock, ProtectHostname and ProtectKernelLogs. It logs them and carries on, so CentOS 7 (systemd 219, which install.sh explicitly supports) runs with less hardening than the unit lists. install.sh and update.sh now print which protections need a newer systemd, which ones still apply, and that upgrading systemd is what changes it. - the updater's writability guard asked [[ -w ]] about the parent directories. It creates and removes a probe file instead, so an immutable attribute or a full filesystem is caught as well (a read-only mount was already caught). - the generator's comment claimed to resolve the folders the service actually uses, while the shipped unit hard-codes WorkingDirectory= and ExecStart= under /usr/local/x-ui. The comment now states what XUI_MAIN_FOLDER really feeds -- the location install.sh/update.sh install into and the base for a relative XUI_BIN_FOLDER -- and that a relocated main folder needs the unit edited too. Rejected from the same review, with the evidence: that [[ -w ]] cannot see a read-only mount (access(W_OK)/faccessat consults __mnt_is_readonly before the mode bits), and that the /etc ReadWritePaths entry is an exception granted for /etc rather than a default store already in the list. Harness extended: 19 checks (escaped %, the old-systemd note, whitespace and duplicate folders, seccomp gating, the read-only guard) plus bash -n on the three scripts, all passing. * fix(systemd): name the hardening old systemd really ignores The old-systemd note fired only below 239 and listed wrong versions: RHEL 8 (239) and Debian 10 (241) silently lose ProtectHostname and RestrictSUIDSGID (242), ProtectKernelLogs (244) and ProtectClock (245) with no note, while CentOS 7 was told NoNewPrivileges (187) and ProtectHome=read-only (214) were not applied although both are. The note is now built from a directive/version table taken from systemd.exec(5) and lists only what the running systemd lacks. Also drop the removal of 10-xui-write-paths.conf: only an intermediate commit of this branch wrote that file, no release ever shipped it. --------- Co-authored-by: Кот <kot@zeroclaw.local> Co-authored-by: Sanaei <ho3ein.sanaei@gmail.com>
2026-10-05 09:47:10 -07:00
if ! _xui_service_write_paths_dropin; then
echo -e "${yellow}Warning: could not refresh ${xui_service}/x-ui.service.d/10-xui-sandbox.conf.${plain}"
echo -e "${yellow}If XUI_DB_FOLDER or XUI_LOG_FOLDER points outside /etc/x-ui and /var/log/x-ui, the panel may not be able to write to it under ProtectSystem=full.${plain}"
fi
_xui_warn_unsupported_hardening
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
return 0
}
feat(amneziawg): add native AmneziaWG protocol support (#6105) * feat(amneziawg): add native AmneziaWG protocol backend AmneziaWG (WireGuard plus DPI-resistant obfuscation) needs no Docker here — it runs as a genuine kernel interface via awg-quick/awg, managed the same way internal/mtproto manages mtg: one Inbound row is one desired Instance, and a Manager reconciles running interfaces toward the database every 10s (internal/web/job/amneziawg_job.go) plus immediately after a client edit (applyLocalAmneziaWG). Clients reuse model.Client verbatim (the same PrivateKey/PublicKey/ PreSharedKey/AllowedIPs fields WireGuard already uses), so bulk operations, the QR/share-link modal and subscriptions come from the shared inbound infrastructure instead of a parallel implementation. internal/amneziawg owns the obfuscation param generator/validator (ported from coinman-dev/3ax-ui, upgraded to AmneziaWG 2.0's S3/S4 padding and I1 signature packet) and the exec wrapper around awg-quick/awg, with fingerprint-based reconcile (noop / reload-via- syncconf / full restart) mirroring mtproto.Manager so a same-protocol edit doesn't force an unnecessary interface bounce that would drop every peer's connection. Frontend and install.sh's DKMS/awg-tools setup are tracked separately; this is backend-only. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): add frontend support and fix a Go->Zod generator gap Wires the amneziawg protocol through the panel UI the same way every other protocol is registered: a Zod settings schema (nested {server, clients}, matching the Go JSON exactly), the protocol enum, the inbound-form's per-protocol fields component and its tab-visibility allowlist, the default-settings factory, the client schema dispatcher, and the sniffing-capability exclusion (no Xray inbound exists for amneziawg, same as mtproto). Client key/allowedIPs fields are reused rather than duplicated: since AmneziaWG clients are wire-identical to WireGuard clients (same model.Client fields), ClientFormModal renders one shared field block for both, switching only the visible label by which protocol is active. The private-key input also gets a live public-key sync via a new useEffect, because unlike WireGuard's Xray-native inbound (which re-derives its public key at runtime and never stores one), AmneziaWG's server.publicKey is a real persisted field the Go backend reads directly — free-typing a new private key without this would silently save a mismatched keypair. Adds a downloadable per-client .conf (amneziawgConfig.ts, mirroring wireguardConfig.ts) with the obfuscation lines, and an InboundOption.AwgServer field on the Go side so the config builder gets the full server block in one round trip. Along the way, running tools/openapigen surfaced a real bug: it doesn't flatten anonymously-embedded Go structs the way encoding/json does, so ServerSettings embedding Obfuscation20 produced a Zod schema with a nested `obfuscation20` key that never matches the real wire JSON. Fixed by un-embedding (flat fields + an accessor method) and registering internal/amneziawg in the generator's own package list, which had been silently emitting a dangling schema reference. English and Russian translations are complete; the other 10 locale files still fall back to English for the new keys. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): complete frontend parity for the Inbounds list page The Clients page (form, CRUD, QR/config) already worked from the prior commit; this closes the remaining gap on the Inbounds side and in a couple of protocol allowlists that a plain search for existing wireguard/mtproto handling turned up. lib/xray/inbound-link.ts gets amneziawg-specific link/config builders (genAmneziaWGLink/genAmneziaWGConfig, plus the *s fan-out variants) mirroring the wireguard ones — AmneziaWG has no legacy peers-array to fall back to, so these read settings.clients directly and add the obfuscation lines every client must share with the server. Wired into genInboundLinks generically, and into three consumers that call the wireguard builders directly rather than through that dispatcher: QrCodeModal, InboundInfoModal, and InboundsPage's bulk export. ClientInfoModal, ClientBulkAddModal, and the bulk attach/detach modals each had their own protocol allowlist that needed amneziawg added alongside wireguard/mtproto. Two real gaps surfaced by grepping every remaining 'wireguard' / Protocols.WIREGUARD hit in frontend/src rather than trusting the checklist was exhaustive: - useInbounds.ts's TRACKED_PROTOCOLS gates the deactive/depleted/ expiring/online client counts shown per inbound on the list page; without amneziawg those counts would silently read zero. - inbound-tag.ts is an explicit client-side mirror of the Go backend's port_conflict.go (the file says so itself: "Keep in sync"). It still only special-cased wireguard for UDP, so an amneziawg inbound would have fallen through to the TCP default and disagreed with the backend's own port-conflict math. Also finishes translating the AmneziaWG UI strings into the 11 locale files that were still falling back to English (ar-EG, es-ES, fa-IR, id-ID, ja-JP, pt-BR, tr-TR, uk-UA, vi-VN, zh-CN, zh-TW), matching en-US/ru-RU key-for-key (26 new keys, verified by count in every file). Not run anywhere: npm run typecheck / build. This machine has neither Node nor npm, so nothing here has compiled — reviewed by hand plus brace/paren balance checks and cross-referencing the generated Zod/TS types. Treat this as needing a real typecheck before shipping. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * docs(install): note that AmneziaWG kernel module install is still manual Tracked separately (not yet ported into this script) — see coinman-dev/3ax-ui's install_amneziawg for the reference approach (ppa:amnezia/ppa). Also serves as a real, path-filter-matching change to get the previous empty commit's CI trigger to actually fire — release.yml's push trigger is paths-scoped and an empty commit changes no files, so it never matched. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): add a button to randomize obfuscation parameters Mirrors the existing key-regenerate button next to the private key field. Client-side randomization matches the ranges/constraints of GenerateObfuscation20's "default" preset (internal/amneziawg/params.go) closely enough for a form suggestion — the user can still hand-edit any field afterward. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(install): auto-install the AmneziaWG DKMS module + amneziawg-tools Ports install_amneziawg from coinman-dev/3ax-ui's install.sh, adapted to this script's broader distro coverage and NONINTERACTIVE convention: - Ubuntu/Debian/Armbian: ppa:amnezia/ppa (primary, tested path), with a reachability pre-check for the Launchpad PPA host — often blocked by hosting providers, especially Russian VPS — so a flaky network skips the feature instead of hanging apt through several retries. - Fedora/RHEL-family, Arch/Manjaro/Parch: best-effort fallback to plain wireguard-tools (+ AUR amneziawg-dkms via yay/paru when available), with a manual-install pointer. - Everything else: manual-install pointer only. Also installs ndppd and persists IPv4/IPv6 forwarding (for the future IPv6/NDP phase, not yet wired into the panel) and adds a Secure Boot warning at the end of the run, since a DKMS-built module is unsigned and won't load while it's enabled — a common trap on cloud VPS images. Never fatal: the panel installs and runs fine either way, an AmneziaWG inbound just won't bring up its tunnel until the module is present. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): resolve all 3 real CI failures (typecheck/lint/codegen) Found by checking the fork's Actions tab after the last two pushes — the release build passed (it doesn't run these checks) but the separate CI workflow caught three real issues: - golangci-lint (noctx): every internal/amneziawg/manager.go exec.Command call is now exec.CommandContext with a 30s timeout, so a hung awg-quick/awg invocation can't block the reconcile job indefinitely (mirrors internal/mtproto/process.go's own CommandContext usage). - tsc --noEmit: frontend/src/schemas/client.ts's hand-maintained InboundOptionSchema (used by the useClients hook, separate from the auto-generated one in generated/) never got an awgServer field added when the AmneziaWG frontend work was done — every read of inbound.awgServer.* in amneziawgConfig.ts was typing as {}. Added AwgServerOptionSchema, nested (not flattened like wg*) to match what amneziawgConfig.ts already expects. Also guarded server.publicKey in inbound-link.ts's genAmneziaWGLink against the schema's optional type. - codegen staleness: frontend/public/openapi.json is produced by a Node script (gen:api) this machine can't run; hand-applied the exact diff the CI failure log already showed (amneziawg protocol enum entry, ServerSettings schema, InboundOption.awgServer, one example payload), verified as valid JSON. Also confirmed independently by this run: install_amneziawg (previous commit) installed and loaded the DKMS module successfully on both amd64 and arm64 CI runners. The two "Deploy Smoke Tests" failures are unrelated to this change — this fork has only ever published the dev-latest pre-release, and GitHub's /releases/latest API deliberately excludes pre-releases, so the smoke test's no-argument install path (which resolves "latest") has nothing to find. Not a regression; needs an actual tagged release whenever that's wanted. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): Phase 2a — IPv6 support + NDP proxy Adds native dual-stack IPv6 to AmneziaWG inbounds, ported from coinman-dev/3ax-ui's approach: - ServerSettings gets ipv6Enabled/ipv6Subnet/ipv6ExternalInterface; Instance carries the server's own IPv6 address (first host of the subnet) alongside its IPv4 one. - defaultAmneziaWGClients allocates an IPv6 host address per client (second AllowedIPs entry) when the server has IPv6 enabled, reusing allocateWireguardAddress — which needed a real fix along the way: it always suffixed "/32" regardless of address family, which is wrong for an IPv6 host address (needs /128). Now family-aware. - generateServerConfig's PostUp/PostDown gains IPv6 forward-accept rules, proxy_ndp sysctl, and one `ip -6 neigh add/del proxy` entry per enabled peer with an IPv6 address — the lightweight per-client method, not the ndppd-daemon whole-subnet method (not worth the config-file-management complexity at this scale; ndppd itself is still installed by install.sh in case that changes later). - ValidateIPv6Subnet rejects a malformed subnet before save. - Frontend: ipv6Enabled/ipv6Subnet/ipv6ExternalInterface fields on the AmneziaWG inbound form, EN+RU translations, openapi.json/generated/* regenerated (the latter via `go run ./tools/openapigen`, pure Go). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): fill in IPv6 fields missed by the Phase 2a commit Two real gaps the CI caught (both new fields, both my miss): - inbound-defaults.ts's createDefaultAmneziawgInboundSettings() built a server object literal predating ipv6Enabled/ipv6Subnet/ ipv6ExternalInterface — AmneziawgServer's inferred type now requires them (zod .default() fields are non-optional post-parse), so this didn't typecheck at all. - openapi.json's ipv6Enabled property was missing the description the real generator attaches (the Go doc comment covering all three IPv6 fields is attached to the first one) — a one-line diff, but git diff --exit-code doesn't care how small. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): Phase 2b — per-client port-forwarding Admins can now set a per-client ForwardedPorts string (e.g. "80, 443, 8000-8100") that gets DNAT'd + FORWARD'd to that peer's tunnel address via iptables rules in PostUp/PostDown, ported and simplified from coinman-dev/3ax-ui's shared/portfwd. Two decisions worth flagging for future readers: - The iptables --comment tag on each rule is awg-fwd-<fnv32a(email)>, not the raw client email. Email is admin/API-supplied free text that ends up embedded in a shell-executed PostUp/PostDown line; a hash can never carry a shell metacharacter through where raw interpolation could. - The reconcile manager gained a third fingerprint (portFwdFP, next to the existing structural/peers ones). `awg syncconf` only touches the WireGuard peer table — it never re-applies PostUp/PostDown iptables rules — so a port-forward-only change has to force a full awg-quick down+up bounce, same as a structural change, rather than the lighter sync a plain peer add/remove can use. Also fixes a real pre-existing bug found while wiring up IPv6 client allocation in the previous commit's spirit: allocateWireguardAddress always suffixed "/32" regardless of address family, which produced invalid host bits for IPv6 (needs "/128"). ForwardedPorts flows through model.Client -> model.ClientRecord (gorm column wg_forwarded_ports, auto-migrated) -> ToRecord/ToClient/ MergeClientRecord, mirroring the awgServer field's earlier lesson that new fields need checking against a second, hand-maintained persistence-layer struct. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): route a client's traffic through Xray via the Routing page Every enabled AmneziaWG inbound gets its own Xray TPROXY bridge automatically, with no toggle to enable first: a loopback dokodemo-door inbound (sockopt.tproxy) tagged with the AmneziaWG inbound's own real tag, so it's already selectable in the existing Routing page's inbound-tag picker — the same trick the mtproto sidecar's own bridge already relies on (InboundService.GetInboundTags is a plain, protocol-blind SELECT over every inbound row's tag, no dedicated UI plumbing needed). internal/amneziawg's defaultPostUpDown TPROXYs every peer's traffic into that bridge unconditionally; the bridge's port is derived deterministically from the inbound's id (EgressPortForInbound) so the kernel-side reconcile loop and the Xray-config generator never need to negotiate a runtime value between them. injectAmneziawgEgress never generates a routing rule itself — whether a client's traffic goes anywhere beyond Xray's default routing is entirely up to whatever rules the admin adds through the existing Routing UI (pick the AmneziaWG inbound's tag as source, optionally a specific peer's IP via that page's own Source-IP field, and an outbound), exactly the same workflow as routing any other protocol. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): recover orphaned interfaces after an ungraceful exit Two gaps left an AmneziaWG interface stuck outside the manager's control after a crash (kill -9/OOM/panic skips StopAll): - ensureRestart's teardown was gated on the in-memory `exists` map, which is always empty on a fresh process, so a survived interface never got interfaceDown before interfaceUp tried `ip link add` against a name the kernel already had — failing forever and never populating m.ifaces, so traffic accounting silently stopped and the inbound could never be removed. Gate on isInterfaceUp instead, which checks real kernel state rather than this process's own bookkeeping. - An inbound deleted from the database entirely while the panel was down has no entry in `desired` ever again, so it never reaches the per-id cleanup loop in Reconcile (which only walks m.ifaces). Add a one-time sweepOrphansLocked scan of configDir, mirroring mtproto.Manager.sweepOrphansLocked, that tears down and removes any leftover interface/config not in the current desired set. Found by the automated review on MHSanaei/3x-ui#6105 (Finding 1). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * i18n(amneziawg): backfill IPv6/obfuscation/port-forwarding keys in 11 locales Only en-US/ru-RU ever got these 9 keys as each AmneziaWG feature landed (the regenerate-obfuscation button, then Phase 2a's IPv6 fields, then Phase 2b's per-client ForwardedPorts) — the other 11 locale files were never backfilled, so i18next has been silently falling back to English for all of them since Phase 1. Cosmetic-only (never broke anything), but now closed for every shipped locale. * fix(amneziawg): resolve 7 Medium findings from the automated PR review Each is independently reproducible; fixed together since one review pass found all of them. - manager.go: the shared "ip rule add fwmark" policy route had no existence check, so it duplicated in "ip rule show" on every interface bounce (which hostRulesFingerprint forces on any client add/remove/ re-IP). Now checked via "ip rule list | grep -q ..." first. (Finding 2) - params.go: ExternalInterface, IPv6ExternalInterface, and subnetIp/ subnetCidr are interpolated unescaped into a shell-executed PostUp/ PostDown line, but only obfuscation and the IPv6 subnet were validated before save. Added ValidateInterfaceName (a strict charset+length pattern) and ValidateSubnetIPv4 (netip.ParsePrefix), wired into normalizeAmneziaWGSettings. (Finding 3) - amneziawg_job.go: IsAwgInstalled() existed but nothing ever called it, so a host without awg/awg-quick (the Docker image, RHEL, Arch, a failed install.sh PPA step) logged a reconcile failure every 10s forever. Now checked once an inbound actually needs it, warning once instead of spamming. (Finding 4) - client_inbound_apply.go: the WireGuard/AmneziaWG credential carry-forward (added so a metadata-only client edit doesn't rotate keys) never covered ForwardedPorts, so a partial edit -- an API call or Telegram-bot toggle that omits the field -- silently wiped a client's port-forwarding spec. Carried forward and written back the same way the key fields already are. (Finding 5) - manager.go: hostRulesFingerprint keyed each peer on its IPv4 address only, and structuralFingerprint omitted IPv6Enabled/IPv6ExternalInterface entirely, so an IPv6-only change could pick the syncconf reload path (which never re-runs PostUp, leaving a stale NDP-proxy entry) or be a complete no-op. Both fingerprints now cover the IPv6 fields. (Finding 6) - port_conflict.go: the AmneziaWG egress bridge (injectAmneziawgEgress) binds 127.0.0.1:63100+id with no collision check anywhere, since it isn't a database row the ordinary port-conflict query can see -- same blind spot the reserved Xray API port already has its own check for. Added the equivalent check for the AmneziaWG bridge port. (Finding 7) - install.sh: install_amneziawg ran unconditionally for every install/ update, building a DKMS kernel module and enabling host-wide IPv4/IPv6 forwarding whether or not the feature is ever used. Gated behind a new should_install_amneziawg (XUI_INSTALL_AMNEZIAWG=true/false, or an interactive y/N prompt defaulting to no). Also replaced the deprecated apt-key adv with a dedicated keyring + signed-by= on the Debian branch, and guarded its sources.list appends against duplication on a retried install. (Finding 8) Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): make the Xray TPROXY bridge a per-inbound opt-in Addresses Finding 10 from the automated PR review: an always-on TPROXY bridge makes every AmneziaWG tunnel hard-depend on Xray being up (all traffic, including DNS, drops whenever Xray restarts), and forces a full awg-quick down+up bounce on any client add/remove/re-IP, permanently losing the syncconf fast path. Adds ServerSettings.RouteThroughXray (off by default): - defaultPostUpDown only emits the TPROXY/policy-route rules when it's on; a plain AmneziaWG tunnel now has zero Xray dependency out of the box. - structuralFingerprint covers it (toggling it changes whether PostUp/ PostDown contain any TPROXY rules at all -- structural, not a per-peer host-rule). hostRulesFingerprint's IPv4 tracking is now itself conditional on RouteThroughXray (and IPv6 tracking on IPv6Enabled), so an instance that never uses either keeps the syncconf fast path for a plain peer re-IP. - injectAmneziawgEgress only creates a bridge for inbounds that opted in; checkAmneziawgEgressConflict (the Finding-7 fix) now parses each candidate through InstanceFromInbound so a non-routed inbound's port is correctly never treated as reserved. - New inbound-level Switch in the AmneziaWG form; the actual outbound decision is still made entirely through the panel's stock Routing page, same as before -- only whether the bridge exists at all is now a choice. Translation keys added to all 13 locales in the same commit this time, not backfilled later (see Finding 9's lesson). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): resolve 4 Low findings from the automated PR review - manager.go: serverAddress assumed subnetIp always ends in ".0"; a base like "10.8.1.5" was used verbatim as the server's own address, eventually colliding with peer allocation (which starts at .2 upward). Now derives the first host of the actual subnetIp/subnetCidr network via netip, matching serverAddressV6's own approach. A /32 base (no host bits at all) is still used as-is. (Finding 12, partial -- the /16 pool-widening half of this finding only exists on the upstream-pr/amneziawg branch's merged client_wireguard.go, not here; handled separately on that branch.) - manager.go: ensureLocked carried the previous per-peer traffic counters (`last`) forward even through a full restart, but awg-quick down+up resets the kernel's own counters to zero -- the next CollectTraffic computed a large negative delta (clamped to 0), silently discarding real traffic. Extracted the decision into nextTrafficBaseline: only a reload (syncconf) preserves the baseline. (Finding 13) - portfwd.go: exported ForwardedPortsInclude; inbound_amneziawg.go's new checkForwardedPortsConflict uses it to reject, at save time, a client's forwardedPorts that would DNAT the panel's own port or another enabled inbound's port to the tunnel client -- portForwardLines has no destination restriction, so this collision was previously silent. Wired into both the single-client update path and the add-client path (client_inbound_apply.go), plus normalizeAmneziaWGSettings for the whole-inbound save path. (Finding 14) - inbound.go: InboundOption.AwgServer sent the whole ServerSettings struct including PrivateKey to GetInboundOptions callers -- a shared, admin-wide dropdown-filling endpoint the frontend's own AwgServerOptionSchema never reads that field from. Redacted it before assigning. (Finding 11) Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): don't widen the peer address pool past AmneziaWG's own subnet Completes Finding 12 from the automated PR review (the serverAddress half of this finding was already fixed on main and cherry-picked here). This half is specific to this branch: allocateWireguardAddress's /16 pool-widening fallback is an independent addition from upstream's own main that this branch inherited during the cherry-pick rebase -- it doesn't exist on the fork's own main at all, so this fix can't be cherry-picked the normal way and is committed directly here. Widening is safe for WireGuard's own Xray-native inbound (AllowedIPs isn't tied to a strict kernel interface subnet), but AmneziaWG's kernel interface Address is exactly the configured subnet -- an address allocated from the containing /16 once the /24 fills up would be silently unroutable. allocateWireguardAddress now takes an explicit allowWidening bool: WireGuard's own caller passes true (unchanged behavior), AmneziaWG's passes false (fails loudly on exhaustion instead). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * docs(docker): note that AmneziaWG doesn't work in this image Investigated: the image is Alpine-based, and AmneziaWG's own packaging (DKMS module + amneziawg-tools) doesn't target Alpine/musl at all -- unlike the Debian/Ubuntu/Fedora/Arch paths install.sh already handles, there's no package to apk add even with full host network/capabilities. The panel already degrades gracefully (IsAwgInstalled() logs one warning instead of retrying forever), so no code change is needed -- just made the reason explicit at the point where a user would reach for cap_add/ network_mode to try to work around it. * fix(sub): include amneziawg inbounds in subscription links getInboundsBySubId's SQL protocol allowlist never had 'amneziawg' added, so every AmneziaWG client was silently excluded from all three subscription formats (plain/individual links, JSON, Clash) and from the Telegram bot's QR/individual-link buttons, which fetch through the same path. genAmneziaWGLink itself was already fully implemented and already wired into GetLink's dispatch switch -- it just never got a chance to run. Same bug shape as the earlier TRACKED_PROTOCOLS frontend gap: a hardcoded protocol list one entry short. Found while investigating whether the Telegram bot needed AmneziaWG- specific client-management code -- it doesn't (the bot itself is fully protocol-agnostic), but this is the actual root cause of "can't share an AmneziaWG client's config via the bot." Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(inbound): enforce node-eligibility server-side, not just in the UI Investigated multi-node interaction with AmneziaWG: the master's own reconcile (DesiredAmneziaWGInstances) and Xray config generation (injectAmneziawgEgress, the GenXrayInboundConfig protocol skip) all correctly filter on NodeID IS NULL, so a node-assigned AmneziaWG (or MTProto) inbound would never be managed by the master. But nothing stopped one from being created that way: NODE_ELIGIBLE_PROTOCOLS (frontend/src/pages/inbounds/form/InboundFormModal.tsx) only hides the node picker client-side -- a direct API call could set nodeId on an AmneziaWG inbound, which every node then reconciles as an ordinary local inbound (nodes run the identical binary, full cron suite included), leaving it running unmanaged and untracked by the master's own AmneziaWG bookkeeping. Added isNodeEligibleProtocol (inbound_protocol.go), mirroring the frontend's allowlist, and enforced it in both AddInbound (the actually exploitable path -- nodeId comes straight from the request) and UpdateInbound (defense in depth; NodeID is already restored from the stored row there before this check, so it mainly guards against a protocol change on an existing node-hosted inbound). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): allow TPROXY-marked traffic through a default-deny INPUT chain TPROXY never rewrites a packet's own destination address, only the routing decision. A default-deny firewall whose INPUT chain sanity-checks "is this destination actually local" (UFW's ufw-not-local, via addrtype --dst-type LOCAL, is a concrete example) silently drops the redirected packet before Xray's socket ever sees it -- RouteThroughXray looked fully configured (TPROXY rule present and counting, Xray listening with IP_TRANSPARENT set) yet every peer's traffic vanished with no trace on either side. Adds an idempotent, never-torn-down "iptables -I INPUT 1 -m mark --mark <fwmark> -j ACCEPT" alongside the existing shared policy route, so this works regardless of which firewall manager owns the rest of the INPUT chain. * fix(frontend): give AmneziaWG the same UDP tag and its own tag color The Inbounds list only special-cased isWireguard/isHysteria for the "UDP" network badge, so an AmneziaWG row showed just the bare protocol tag with no transport badge next to it. Added the missing isAmneziawg flag (mirrors isWireguard exactly) and wired it into the same branch. Client-row protocol-color maps in ClientsPage/HostList had no amneziawg entry, silently falling back to grey -- ClientInfoModal already had amneziawg: 'yellow' from earlier work, these two just never got it. * feat(logs): show which AmneziaWG client an access-log line belongs to The dokodemo-door TPROXY bridge every AmneziaWG peer's traffic is routed through has no per-user identity, so Xray's own access log never carries an "email:" token for these lines -- the Access Logs modal showed a blank Email column for every in-*-udp row, even though every other protocol's rows show the client normally. The peer's decapsulated tunnel IP does survive as the log's "from" address, and that IP deterministically maps to exactly one configured peer. Builds a "<inbound tag>|<ip>" -> email index from the same AmneziaWG inbounds already parsed elsewhere (amneziawg.InstanceFromInbound), and fills in Email from it whenever the raw log line didn't have one. * fix(amneziawg): enable sniffing on the TPROXY bridge Domain-based Routing rules could never match RouteThroughXray traffic: an AmneziaWG peer resolves DNS itself, through the tunnel, before ever sending a packet, so the decapsulated traffic TPROXY hands to the bridge is already a bare destination IP with no domain name attached at the network layer. Every other inbound recovers this via sniffing (confirmed working for the stock wireguard inbound, which does have it configured); the bridge never got a sniffing block at all, so only tag/IP/network-based rules could ever match it -- any domain rule above it in the list was silently unreachable. * docs: add an AmneziaWG config page and list it as a supported protocol Closes the PR checklist gap: the feature shipped with zero mention on the docs site. Mirrors reality.mdx's structure (key settings, setup steps, config excerpt) and notes the Docker/multi-node/Telegram-bot caveats the PR itself is honest about not having confirmed. * fix: address the fresh review round on PR #6105 (8 findings) 1. hostRulesFingerprint didn't account for ForwardedPorts when RouteThroughXray was off, so re-IPing a peer with port-forwarding configured left stale DNAT rules pointing at an address the next peer could be handed. 2. Server/client config values (keys, email, I1) were never validated for control characters before being written into the generated .conf; a newline could smuggle a PostUp hook into awg-quick's parser. Added ValidateConfigValue at save time and a sanitizeConfigValue backstop at render time. 3. checkForwardedPortsConflict didn't scope to node_id IS NULL, so a port used only on a different node produced a false collision; also hoisted the panel-port/inbounds lookup out of the per-client loop (portConflictContext) so N clients cost one query, not N. 4. PostDown commands were ";"-joined and abort on the first failure; appendOrTrue makes teardown best-effort so an external firewall flush can't leave DNAT rules to accumulate across bounces. 5. The "ip rule list | grep -q" existence check could SIGPIPE under pipefail and re-add a duplicate rule; switched to grep -c >/dev/null. 6. Ported the vpn:// share-link format (base64url of the plain .conf text, matching the real AmneziaVPN app) onto this branch -- it had only ever landed on our own fork's main, so this PR branch was still on the old amneziawg://+query-params scheme our own docs no longer described. Also corrected the docs' install.sh claim (opt-in/ interactive, not automatic) and stale pre-opt-in comments in route_egress.go. 7. install.sh: Arch's ndppd install used pacman -Syu (full system upgrade) instead of -Sy like every other call in the script; and should_install_amneziawg re-prompted on every `x-ui update` even when awg was already installed. 8. CollectTraffic could clobber a concurrent restart's freshly-reset (empty) traffic baseline with stale pre-restart counters, since getPeerStats runs lock-free; now checks pointer identity before writing back. sweepOrphansLocked permanently disabled itself on a transient os.ReadDir failure instead of allowing a retry. go build/vet/test and frontend typecheck/lint/build/vitest all pass. * fix(install.sh): check the live sysctl value, not sysctl.conf text Reviewer feedback (cherts, PR #6105): grepping /etc/sysctl.conf for the setting name is unreliable -- many distros split sysctl config across /etc/sysctl.d/*.conf, and /etc/sysctl.conf can be a symlink into that directory, so the check can miss an already-active setting (harmless duplicate append) or match a disabled/commented line (forwarding silently stays off). Query the live value via `sysctl -n` instead, which is accurate regardless of which file set it. Applied the same fix to both the IPv6 and IPv4 checks for consistency. * fix: update inbound_amneziawg.go to the split buildInboundForLocalRuntime Same fork-only-file blind spot as the one caught on our own main after the 3.6.0 sync: upstream split buildRuntimeInboundForAPI into buildInboundForNodePush / buildInboundForLocalRuntime (part of the node-sync client-deletion fix, 5bc81dfd), updating every call site it could see. This file doesn't exist upstream, so it kept calling the old name even after the branch merged in that commit. * fix(frontend): recognize AmneziaWG's vpn:// scheme in share-link labels The shared link-tag/label helper (used by the client info modal, QR modal, and subscription page) had no entry for the vpn:// scheme AmneziaWG links use, so it fell through to the generic fallback: a plain "Vpn" tag with no color, and an empty remark/port that made the row's title fall back to "Link N" instead of the inbound's actual name:port — unlike every other protocol, which shows its real tag and label. vpn:// links are base64url of a plain .conf text (matching the real AmneziaVPN app's own share-link format), not a structured URL, so there's no query string or #hash to read a remark/port from. Decode the payload and pull the remark/endpoint back out of the .conf text directly instead. * fix(xray): force a full restart for TPROXY inbounds, never hot-add them Real incident: an AmneziaWG inbound with RouteThroughXray enabled lost all internet on that connection after a migration. Root-caused on the live box -- iptables TPROXY counters were incrementing (packets correctly redirected to 127.0.0.1:63110), but nothing was actually listening there (ss showed nothing on that port) until a full `systemctl restart x-ui`, after which the bridge came up immediately. Xray-core's gRPC AddInbound reports success for a new sockopt.tproxy inbound (internal/amneziawg's own Xray egress bridge is the only kind this fork ever generates) but doesn't reliably bind a working listener for it outside of process startup -- the bridge silently never comes up, and RouteThroughXray traffic goes nowhere until the next full restart happens to occur for an unrelated reason. diffInbounds already has this exact defensive pattern for REALITY inbounds ("a gRPC remove+add does not reliably rebuild the REALITY authenticator"), just never extended to TPROXY, and only in the already-existing-then-changed branch -- the "brand new inbound" branch had no such guard at all, which is exactly the path a freshly-enabled RouteThroughXray bridge takes. Added inboundUsesTproxy and wired it into both branches. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): flag Xray for resync when a peer edit changes qualifying state Real production bug, root-caused on iiadmin-vps: updateAmneziaWGInbound/ AddInbound/DelInbound only ever updated the kernel interface via amneziawg.GetManager() -- they never called SetNeedRestart the way every other protocol's mutation path does (client_crud.go, inbound.go, etc. all do). injectAmneziawgEgress's TPROXY bridge inbound depends on InstanceFromInbound finding at least one qualifying peer plus RouteThroughXray, so an edit that flips that (first peer added, last one removed, RouteThroughXray toggled on) previously required a full panel restart before the bridge actually got created, with no error anywhere: the kernel interface would handshake fine, but traffic redirected into the bridge's TPROXY port went nowhere because nothing was listening there. diffInbounds/inboundUsesTproxy already correctly force a full restart for a brand new TPROXY inbound (bdee0a20) -- that part was never the bug. The gap was entirely upstream: nothing ever told Xray a resync was even needed. * fix(clients): reject AllowedIPs already used on another WireGuard/AmneziaWG inbound defaultWireguardClients/defaultAmneziaWGClients only ever checked uniqueness against their own inbound's client list, so two inbounds sharing a subnet (same protocol or not) could silently hand out or accept the same address -- the exact scenario behind a real duplicate-IP incident where a WireGuard and an AmneziaWG client both ended up on the same address. otherTunnelAllowedIPs now collects every address already claimed on every other tunnel inbound and folds it into both the auto-allocation pool and the manual-entry collision check, naming the other inbound in the error when it fires. * fix(frontend): add the missing AmneziaWG config download on the sub page The subscription page already gave WireGuard links their own "Config" block (copy/download/QR of the actual .conf, via wireguardConfigFromLink reversing the wireguard:// query params) but had no equivalent for AmneziaWG's vpn:// links -- its isWireguardLink gate never matched them, and no reverse-parse helper existed for this page specifically. Every other surface (InboundInfoModal, ClientInfoModal, ClientQrModal) already had this parity; this was the one page that didn't. Fixed by adding amneziawgConfigFromLink (inbound-link.ts), simpler than its WireGuard counterpart since a vpn:// payload already *is* the plain .conf text -- just base64url-decode it, no query-param reconstruction needed -- and wiring it into SubPage.tsx alongside the existing WireGuard block, reusing the same pages.clients.amneziaWgConfig label the other three surfaces already use. * fix(xray): force a full restart for password-auth SOCKS5 hot-apply Real production incident: editing a client under an AmneziaWG inbound left its embedded SOCKS5 relay's settings byte-different (a new account list), and Xray's gRPC remove+add hot swap silently dropped the account for a peer whose email contained non-ASCII characters -- its tunnel kept handshaking fine but all its traffic got rejected at the SOCKS5 layer, while every other peer on the same relay was unaffected. A full restart (reading the same JSON straight from disk) always produced the correct account list. socks isn't in userDiffableProtocols (that only covers vless/vmess/trojan's clients+email shape, not accounts+user), so any settings drift on this inbound fell through to the generic remove+add path. Forces a restart instead, the same defensive choice already made for REALITY and TPROXY -- scoped to auth:"password" specifically so the other, noauth SOCKS5 bridges (panel/node/mtproto egress) keep the cheaper hot path. * Fix Attach reusing one identity's address across wg/awg inbounds ClientService.Attach deliberately copies one identity's stored AllowedIPs into every WireGuard/AmneziaWG inbound it's attached to in the same call, so the same person gets the same tunnel address on every protocol they use. Its loop calls addInboundClient once per inbound, and each of those independently computes otherTunnelAllowedIPs -- so by the second inbound in the batch, the first inbound's just-written copy of this identity's own address looked like a cross-inbound collision against itself. Real production symptom this caused: detaching then re-attaching a client to both wg and awg failed with "wireguard: allowedIPs entry X is already used by a client on inbound 'awg' (#N)" -- the exact address the identity is supposed to keep, rejected as if it belonged to someone else. Add a selfEmails exclusion to otherTunnelAllowedIPs and populate it from the client(s) being processed at the one real call site. Safe unconditionally: ClientRecord.Email is globally unique, so a match can only ever be this same identity's own entry on a sibling inbound, never a genuine different client's address. Reproduced the underlying mechanism live (manual entry correctly rejected as a cross-inbound collision; fresh auto-allocation correctly avoided a used address) before writing the fix, to confirm the guard itself works and the bug is specifically in how Attach's per-inbound calls interact with it. * Attach: allocate fresh when re-attaching with no active tunnel The previous fix (82cc69f5) made Attach's own address-reuse correctly not collide with itself across inbounds -- but it still always reused an identity's stored AllowedIPs verbatim, even when that identity currently has zero WireGuard/AmneziaWG attachments at all. A real report from testing this live: an identity fully detached from both its wg and awg inbounds, then re-attached, got its old address back even though several lower addresses were free -- because nothing about being fully detached ever cleared the stored value Attach copies from. Add hasTunnelAttachment, checked once against the identity's CURRENT inbound set before Attach's loop runs: if none of its current inbounds is WireGuard/AmneziaWG, clear the stored AllowedIPs so this attach allocates fresh (matching what a brand-new client would get) instead of resurrecting an address nothing reserves anymore. Left alone when the identity already has an active tunnel elsewhere, so extending it to a second protocol still keeps a consistent address. * Fix TestOtherTunnelAllowedIPsExcludesSelfEmail's own test setup CI caught this: the "genuinely different client" (other@wg) was seeded onto the SAME inbound passed as excludeID, which otherTunnelAllowedIPs already excludes entirely regardless of the selfEmails fix -- so the assertion that its address is still reported could never have passed, proving nothing either way. Move it onto the sibling inbound alongside shared@id, which is what the test actually needs to exercise (two clients on one sibling, one excluded by email, one not). * Attach: never inherit an address that doesn't fit the target inbound hasTunnelAttachment (from the earlier fix, commit 51067f16) only asked "does this identity have ANY tunnel attachment", treating that as license to reuse its stored address verbatim on every inbound being attached. Real production case this missed: an identity's stored address came from WireGuard's own fallback subnet (10.0.0.0/24, used when that inbound has no other clients to infer a base from), then got attached to a second, AmneziaWG inbound configured for a completely different subnet (10.8.1.0/24). defaultAmneziaWGClients's already-set-AllowedIPs branch only checks for collisions, never subnet membership, so the mismatched address was accepted silently -- producing a peer that can never actually connect, since an AmneziaWG address must fall inside the kernel interface's own configured subnet to be routable at all. Add addressesFitAmneziaWGInbound, checked per inbound inside Attach's loop: if the inherited address doesn't fit the SPECIFIC inbound being attached, clear it just for that one so it gets a fresh, valid allocation instead, while other already-attached inbounds keep their existing values. WireGuard has no equivalent strict subnet requirement (allocateWireguardAddress can widen to a fallback pool for it), so this only ever constrains AmneziaWG targets. * Give WireGuard an explicit, admin-configurable subnet field WireGuard previously had no configurable subnet at all -- only an implicit one, either inferred from existing clients' own addresses (wireguardAllocationBase) or a hardcoded 10.0.0.0/24 fallback when none exist yet. AmneziaWG, by contrast, has always had a real server.subnetIp/subnetCidr field in its settings, editable in the UI. User request: give WireGuard the same treatment. Backend: explicitWireguardSubnetBase reads an optional subnetIp/ subnetCidr pair from the inbound's own settings JSON (mirroring AmneziaWG's defaultAmneziaWGSubnetBases). defaultWireguardClients checks it first; only when unset does it fall back to today's inference-from-existing-clients behavior, so an inbound saved before this field existed keeps working exactly as it always has. Frontend: subnetIp/subnetCidr added to WireguardInboundSettingsSchema and the inbound form (mirroring AmneziaWG's own field layout/labels), with a real default (10.0.0.0/24, the same value the backend already fell back to) seeded for newly created inbounds so the field starts populated and editable rather than blank. Translated across all 13 locales. This also structurally closes the class of bug fixed in 82cc69f5/291c47b3: with wg and awg subnets explicit and independently controllable, an admin who wants matching addresses across both protocols can configure them to actually agree, instead of one silently inheriting the other's incompatible range. * Split the client edit form's AllowedIPs into per-protocol fields A client attached to both WireGuard and AmneziaWG shared one AllowedIPs form field with a dynamically-switching label, so its two genuinely different addresses could never both be shown or edited correctly. Worse, Update/Create broadcast that one shared value to every attached wg/awg inbound with no subnet-fit check, so an ordinary edit save could silently overwrite one protocol's address with the other's -- the same bug class already fixed for Attach, but reachable from any client edit. model.Client gains an optional AllowedIPsByInbound map so a caller can send distinct values per inbound; Update/Create honor it and, when it's absent, clear a shared value that doesn't fit an AmneziaWG inbound's own subnet instead of writing it through. A new TunnelAllowedIPsByInbound read path feeds the real per-inbound address to the client edit form via GET, which now renders two separate, correctly-labeled fields whenever both protocols are attached (unchanged single dynamic field otherwise). * Regenerate openapi.json for the new allowedIPsByInbound field Follow-up to 878ee839: gen:zod (frontend/src/generated) was already regenerated and committed, but gen:api (frontend/public/openapi.json) wasn't, so CI's codegen drift check failed. * Fix build breakage from merging upstream main: Update() gained a limitHwid param Two of our own AllowedIPs tests (not present upstream, so the merge never flagged them as conflicting) still called the old 3-arg Update(inboundSvc, id, client) -- upstream's hardware-ID-limit feature added a required limitHwid parameter that every other caller in this package already passes. Also drop createDefaultInboundSettings from InboundsPage.tsx: the merge conflict resolution kept the import, but upstream's clone-payload refactor (buildClonePayload, inbound-clone.ts) already calls it internally now -- this file doesn't need it directly anymore. * Fix real bug: AmneziaWG clients rejected as "empty client ID" in 3 places Three switch statements on inbound.Protocol handle "wireguard" explicitly (checking client.PublicKey) but fall through to the default case for "amneziawg" (checking client.ID, which AmneziaWG clients never set -- they use PublicKey/Email like WireGuard, not the VMess/VLESS UUID field). This is what the 4 AllowedIPs tests were actually catching: UpdateInboundClient's newClientId derivation hit this same default branch, so every Update() on an AmneziaWG client returned "empty client ID" before ever reaching the AllowedIPs logic being tested. Fixed by adding "amneziawg" alongside "wireguard" in each switch: addInboundClient's per-client validation, UpdateInboundClient's newClientId derivation, and AddInbound's per-client validation (the third one wasn't hit by these tests, but has the identical bug -- creating a brand-new AmneziaWG inbound with a client attached would fail the same way). * refactor(amneziawg): rename Obfuscation20 to Obfuscation31, drop the dead mobile preset Mechanical rename ahead of the AmneziaWG 3.1 parameter work: the type, generator and prose all said 2.0, and the "mobile" generator preset was reachable only from its own test. No behavior change. * feat(amneziawg): AmneziaWG 3.1 obfuscation parameters (backend + generated schemas) Adds the 3.1 parameter surface to the inbound settings and both Go config emitters: I2-I5 signature packets, HeaderProtectionKey (base64 32-byte, shared server<->client), ContentPaddingAddition, the five handshake-timing randomization ranges (RekeyAfterTime/RekeyTimeout/RejectAfterTime/ KeepaliveTimeout/MaxHandshakeAttempts), and the RandomTrailers/ DisableCookies switches. Freshly generated sets fill everything except I2-I5 (matching Amnezia's own generator) with jittered ranges bracketing WireGuard's stock timing constants; every reject window starts >= 30s above the rekey window by construction. Empty fields stay off the wire, so blanking a field disables just that feature. Validation generalizes the H1-H4 range checker for the new uint32-range fields, requires min 1 on timers, cross-checks rekey-vs-reject, and demands a real 32-byte base64 header-protection key. The manager warns once per process when the installed awg tools predate 3.1 but an inbound uses 3.1 parameters (awg-quick rejects unknown keys with a generic error otherwise); apply still proceeds. Requires amneziawg-tools v3.1.20260812+ / module or amneziawg-go v3.1.20260814+ on the host. * feat(amneziawg): emit and randomize 3.1 parameters in the frontend Both client-config emitters (the vpn:// link builder and the clients-page .conf builder) now carry the 3.1 [Interface] lines in the same order as the Go emitters. The obfuscation randomizer moves out of InboundFormModal into a shared lib/xray/amneziawg-obfuscation.ts that also fills the new fields, and createDefaultAmneziawgInboundSettings switches from static values to that generator — a fresh inbound now really gets the unique fingerprint the docs promise instead of the same jc=5/jmin=10 set on every install. Schema parse-time defaults for the new fields stay ''/false on purpose: real values come only from the generator, so resaving an inbound never mutates its stored parameters. A new parity test pins the hand-written AmneziawgServerSchema to the generated ServerSettings key set, so a field added on one side can no longer silently vanish from configs. * feat(amneziawg): 3.1 form fields and translations Inbound form gains inputs for I2-I5, HeaderProtectionKey (filled by the existing obfuscation Regenerate button), ContentPaddingAddition, the five timing ranges, and the RandomTrailers/DisableCookies switches; the MTU input picks up the min=1 its schema already enforced. All 13 locales get the 19 new keys and drop the "2.0" branding from the s3/s4/i1 labels. * docs(amneziawg): document 3.1 parameters; install.sh kernel/version notes The AmneziaWG page's obfuscation section moves from the 2.0 to the 3.1 parameter set: table rows for I2-I5, HeaderProtectionKey, ContentPaddingAddition, the timing-randomization ranges and the RandomTrailers/DisableCookies switches, a requirements callout (tools v3.1.20260812+, module/awg-go v3.1.20260814+, Linux 6.7+ for the DKMS path), and a sample client .conf that matches what the panel actually emits (including the DNS defaults and PersistentKeepalive it always had). install.sh warns before a DKMS build on a pre-6.7 kernel and after any install that left pre-3.1 amneziawg-tools on PATH. Also updates the hosts API operation paths ({id} -> {groupId}) in the stale ru/zh/fa reference pages: syncing docs/public/openapi.json for the new AmneziaWG schema fields surfaced that rename, which had never been copied over, and the docs build fails on paths missing from the spec. * fix(amneziawg): reject control characters and canonicalize 3.1 range values Adversarial review of the 3.1 work surfaced a validation gap: base64.DecodeString silently ignores CR/LF, so a header-protection key that picked up a line wrap in transit decoded to a valid 32 bytes, passed validation, and was emitted verbatim into every client config — where the orphan second line breaks the import while the server (whose emitter strips control chars) keeps running with the correct key. The key and range validators now reject control characters outright. Also from the same review: range values are canonicalized on save ("110 - 140" -> "110-140", whitespace-only collapses to feature-off, closing a case where the server conf rendered an invalid blank-value line the client emitters omitted); the rekey/reject invariant is now enforced against WireGuard's 120s/180s defaults when only one side is set; and the structural fingerprint joins on "\n" instead of "|", which is a legal I1-I5 character and made adjacent free-text fields join-ambiguous. * fix(install): resolve latest release tag via web redirect to dodge API rate limits The non-interactive install smoke test resolved the release version through the unauthenticated GitHub API (api.github.com/.../releases/latest), which allows only 60 requests/hour per IP. The test installs twice in one run, and on shared CI runner IPs the second call gets rate-limited, returns no tag_name, and install.sh treats an empty version as fatal (exit 1) — the same "Failed to fetch x-ui version" real users hit behind CGNAT/shared addresses. resolve_latest_tag() now reads the tag from the github.com releases/latest web redirect (not subject to the API rate limit), falling back to the API only if the redirect yields nothing. Verified with the real deploy/test/smoke-noninteractive.sh (two installs, both green). * fix(amneziawg): three review findings on #6105, plus a comment trim 1. A peer's allowedIPs reached the generated .conf unvalidated and unsanitized, unlike email/publicKey/preSharedKey which normalizeAmneziaWGSettings already guards. A newline in an entry let a following "[Interface]" re-open the interface section, whose "PostUp = ..." awg-quick then runs as root on the next apply. Reproduced end to end against generateServerConfig. The save path now rejects and canonicalizes through normalizeWireguardAllowedIPs, and the render path sanitizes as a backstop for rows predating the validation (an upgrade, a restored backup, a direct DB edit). H1-H4 get the same render-time sanitize, and the two NIC name fields a plausibility check, since stripping control characters alone would still let a shell metacharacter into a root-executed PostUp line. 2. EgressPortForInbound is 63100 + inbound id, so an id past 2435 derives a port above 65535 -- and Xray rejects the whole generated config over one invalid port, taking every other protocol down with it. It now reports ok=false past the range, and both the Xray bridge and its TPROXY rules are skipped instead of emitting an impossible port. 3. The downloadable AmneziaWG .conf read ClientRecord.allowedIPs, a single shared column that holds the WireGuard address for an identity attached to both protocols -- the exact ambiguity tunnelAllowedIPs was added to resolve for the edit form. The info and QR modals already hydrate that field, so they now pass this inbound's own address to the builder. Also trims the comment blocks in the files touched here to the 2-line guidance in CLAUDE.md: internal/amneziawg alone carried 423 comment lines in over-long blocks against 118 for the comparable internal/mtproto, and is now at 110. Every non-obvious constraint is kept (the kernel S1/S2 rule, why PostDown is best-effort, why grep -c and not -q, why the fingerprints split three ways); the narration is gone. Two hot_diff.go comments pointed at an internal/amneziawgnet package and an injectAmneziawgnetSocks function that exist nowhere in the tree; the checks themselves are unchanged. * feat(logs): add an AmneziaWG log view to the overview The overview has an access-log view for Xray but nothing for AmneziaWG, so when a tunnel misbehaves there is no way to see it from the panel at all. A kernel tunnel logs no per-request lines, so the equivalent view is built from the two things it does expose: - Live per-peer activity from `awg show <iface> dump`, joined to the client email through the desired peer set: last handshake, endpoint, allowed IPs, cumulative transfer and online state, newest handshake first. - The panel's own AmneziaWG event lines (interface up/down, awg-quick failures, the pre-3.1 tools warning), which are what actually explain a peer being absent from the table. POST /panel/api/server/amneziawglogs/:count serves both, with the same count + filter contract GetXrayLogs uses, and the modal mirrors XrayLogModal's toolbar, auto-update, mobile cards and download. The action-bar button is gated on a new status.amneziawg.configured, which stays true while an inbound exists but its interface is down -- exactly when the event lines matter. Verified against a running panel: the endpoint returns the peer table and real event lines ("awg/awg-quick not found on PATH", "create config dir: permission denied"), and count and filter both narrow as documented. One of those lines surfaced a Debugf that had been rendering as "for inbound1:amneziawg:"; fixed here since it is now user-visible. * fix(amneziawg): stop double-counting a routed inbound's traffic injectAmneziawgEgress tags its Xray bridge with the AmneziaWG inbound's own tag, so the stock Routing page can target it. Xray therefore reports that bridge's bytes under the inbound's tag, and XrayTrafficJob feeds them to AddTraffic -- which accumulates -- on top of the same bytes AmneziaWGJob already reported from `awg show dump`. An inbound with routeThroughXray on counted roughly twice its real traffic, which also inflates the quota checks that read the same counters. The awg counters are the complete measure: every peer, whether or not TPROXY routed it, and the same wire bytes the per-client totals are built from, so they stay and the Xray rows are dropped. Per-client stats were never affected -- a dokodemo-door bridge has no per-user identity, so Xray emits no user>>>email rows for it. Filtering happens before every consumer, so the DB totals, the external traffic inform and the dashboard's live speed all read one source per inbound. The set of bridge tags now comes from a predicate shared with injectAmneziawgEgress itself, with a test that pins the two together -- naming one tag too few doubles the traffic again, one too many makes real traffic vanish. * fix(amneziawg): align the three .conf emitters on one peer field order The panel builds an AmneziaWG client .conf in three independent places, and they disagreed: buildAmneziaWGClientConfig put PresharedKey right after PublicKey (wg-quick(8)'s own order, and what both WireGuard emitters on the clients side already use), while genAmneziaWGConfig and the Go amneziaWGConfigText put it after Endpoint. A user comparing a subscription link against a downloaded .conf sees the difference immediately, and the generators are exactly the kind of parallel implementation CLAUDE.md warns about drifting. Moves the two outliers onto the wg-quick order. Also drops the stray trailing newline that only appeared when PersistentKeepalive was set, so a config now always ends on its last set field whichever that is -- the same shape all three emitters produce for the same client. Parsing is unaffected either way (the format is order-insensitive, and the AmneziaVPN app reads it as a flat key-value bag), so this changes only the rendered text. Adds a test on each side that pins the peer block's field order, since nothing previously asserted it. * refactor(amneziawg): switch to the embedded amneziawg-go/gVisor architecture Replaces the kernel-module (DKMS) + awg-quick + TPROXY backend with the fork's own embedded design: amneziawg-go runs in-process over a userspace gVisor netstack, and each peer's decapsulated traffic relays into its own loopback Xray SOCKS5 inbound, so Xray's native stats/sniffing/routing work for free instead of through hand-rolled bridges. No kernel module, no DKMS, no Secure Boot conflicts, works the same in a container as on bare metal. - internal/amneziawgnet: new package (Device/UAPI, gVisor netstack, TCP/UDP forwarding, SOCKS5 relay, peer identity, IPv6 host-alias egress identity, per-client port-forwarding) - amneziawg-go v3.1.20260814 + gvisor. - internal/amneziawg: keep the reusable protocol-shape types/validation (Instance/Peer/Obfuscation, InstanceFromInbound); drop the OS-shellout half (awg-quick, TPROXY policy routing, NDP proxy, peer-stats parsing). - internal/web/service: rewire the 5 integration points (job, runtime, client-apply, web shutdown, xray config) from the old manager to the new one; the AmneziaWG log view is rebuilt on the embedded Device's own UAPI dump (extended to carry endpoint/AllowedIPs) instead of `awg show dump`. - install.sh: drop DKMS/ndppd/TPROXY/Secure-Boot installer code (~250 lines) - an entire recurring class of installer fragility goes away. - frontend: drop the now-meaningless routeThroughXray toggle (the relay is always on); keep the field in the Zod schema, unexposed, so it isn't silently stripped from stored settings on next save - two regression tests deliberately depend on the Go struct still carrying it. - docs/i18n: rewrite amneziawg.mdx for the new architecture; drop the dead routeThroughXray translation keys across all 13 locales. Real production throughput (embedded core datapath, isolated bench, same box the kernel-module path was measured on): ~296 Mbit/s up, ~640 Mbit/s down, vs. 414.69 MB/s (~3.3 Gbit/s) for the kernel module on the same hardware - a real gap, tempered by this being single-stream/no-SOCKS5-hop and most VPN traffic being latency-bound rather than throughput-saturating. * fix(amneziawg): restore the branch's own Obfuscation31 shape + 2 CodeQL findings The previous push's wholesale-copy of types.go/params.go from the fork's main branch pulled in that branch's own independent (and incompatible) naming for the same AWG 3.1 feature set: Obfuscation20/GenerateObfuscation20 instead of this branch's already-shipped Obfuscation31/GenerateObfuscation31, and a missing CanonicalizeUintRange -- broke every Go CI job (the whole matrix fails to compile when any one package doesn't, which is why govulncheck/ golangci/postgres-durable-first/race all failed identically, not just go-test). Restores params.go/params_test.go verbatim from this branch's own last commit (a strict superset of validation: it already cross-checks rekey vs. reject timing windows, which the copied version never did) and folds the 3.0/3.1 fields (HeaderProtectionKey, ContentPaddingAddition, the 5 timing fields, RandomTrailers/DisableCookies) into Obfuscation31 itself, matching the original struct exactly instead of as separate top-level Instance fields. instance.go, the two amneziawgnet call sites, and 7 amneziawgnet test files updated to match. Also drops the one test (sanitizeConfigValue) that only ever served the retired kernel-module .conf writer -- correctly not ported, so the test testing it shouldn't have been copied either. Also fixes 2 CodeQL findings the same push surfaced: a clamped uint64->int64 conversion for the new log view's live byte counters (server.go), and an unneeded len+len sum feeding a slice pre-size in the v6-egress outbound merge (xray.go) -- append already grows correctly without it. * chore(amneziawg): regenerate frontend schemas for updated doc comments npm run gen was missed after the previous commit's types.go doc-comment edits (Obfuscation20 -> Obfuscation31, ValidateHeaderProtection -> ValidateObfuscation in the prose) -- openapigen bakes those comments into the generated schema's description field, so the committed frontend/src/generated/schemas.ts and openapi.json still had the old wording. codegen's git-diff-exit-code check caught it correctly. * fix(amneziawg): narrow 2 test fixtures that collided with MaxForwardedPorts TestCheckForwardedPortsConflict_CollidesWithEnabledInboundPort and ..._NoCollisionWhenPortsDontOverlap used "8000-8100"/"9000-9100" as their ForwardedPorts fixture -- 101 ports each, one over MaxForwardedPorts (100). The cap check (checkForwardedPortsConflict, added this session alongside the SOCKS-phantom-port check) fires first, so both tests got "more than 100 forwarded ports" instead of ever reaching the collision logic they're actually testing. The cap itself has its own dedicated boundary test already; these two just needed a narrower range that still covers/misses port 8080 as intended -- 8075-8085 and 9075-9085, 11 ports each. * fix(amneziawg): checkAmneziawgnetSocksConflict had no receiver in its new home My merge-conflict resolution kept this as a method call (s.checkAmneziawgnetSocksConflict) inside checkPortConflictTx, a plain function with no *InboundService receiver -- upstream's #6225 fix moved the port-conflict check out of the (s *InboundService) method and into this new tx-scoped free function, and I didn't notice the call site needed to change shape too. CI caught it immediately (undefined: s); nothing in this specific package can be locally verified past internal/database's own unrelated, pre-existing CGO build issue on this dev machine. Since the signature had to change either way, folded in the fix already flagged as a separate follow-up: checkAmneziawgnetSocksConflict now takes the caller's db handle instead of fetching its own via database.GetDB(), so it actually runs inside the same serialized transaction #6225 introduced -- previously it sat right next to that race fix without benefiting from it. * fix: address the review findings on the embedded AmneziaWG PR 5 blocking findings: - Floor S3/S4 at 12 in both obfuscation generators (Go and frontend) and reject a hand-edited value below that when HeaderProtectionKey is set -- IpcSet requires it, and ~39% of previously-generated sets violated it silently. - Guard PrivateKey/PrimaryDNS/SecondaryDNS/remark against newline injection in the AmneziaWG .conf builder (both the Go subscription-link path and the frontend downloadable-config path) -- unguarded, any of them could inject an arbitrary config line into a subscriber's client. - Bound the derived AmneziaWG SOCKS relay port to <= 65535 once an inbound's id is known, and check the reverse direction (does the relay port collide with an existing inbound's port) on both create and update -- previously only port -> relay collisions were checked, not relay -> port. - Gate injectAmneziawgV6Egress on the same V6AliasesActive predicate desiredV6Aliases already uses, so the two can't disagree about whether a peer's IPv6 identity is actually active at the OS level. 2 minor findings: - Fix the forwarded-ports cap check's off-by-one (a spec covering exactly the cap was rejected as if it were over it). - Correct docker-compose.yml's stale comment describing the retired DKMS/kernel-module architecture. * chore: retrigger CI build (armv5) failed on a transient Go module proxy network error (INTERNAL_ERROR stream reset on sagernet/sing), unrelated to this PR's changes. * docs: fix doc comments still describing the retired DKMS/awg-quick design A few doc comments (and one illustrative test log line) survived the embedded-architecture cutover unchanged and now contradict the code they sit next to: - internal/amneziawg/types.go's package comment claimed this package still owns a Manager that reconciles OS-level interfaces via awg-quick/DKMS -- that Manager was removed; the reconcile loop lives in internal/amneziawgnet now, and this package is protocol-shape-only. - internal/amneziawg/params.go's ValidateObfuscation/ValidateConfigValue comments cited "awg-quick up" / "awg-quick executes as root" as the reason to validate -- the server itself never calls awg-quick in this architecture; the same value still reaches a real rendered .conf that a client app or an admin's own awg-quick CLI applies downstream, so the validation is still warranted, just for a different consumer. Mirrored the same fix in inbound_amneziawg.go's matching comment and its test's comment. - internal/amneziawgnet/manager.go's Manager doc comments (x3) pointed readers at "internal/amneziawg.Manager" for comparison -- that type no longer exists in this diff at all. Repointed at internal/mtproto.Manager, the pattern this was actually modeled on and the one that's still real. - Swapped one test's illustrative "awg-quick up awg2 failed" log line for a message shaped like this architecture's actual amneziawgnet logging, so a reader skimming the test doesn't wonder whether the server still shells out to awg-quick. No behavior change. * fix(docs): re-run codegen for xray-settings.mdx after conflict merge The automated conflict-resolution hand-merge for this generated file was content-correct but didn't byte-match a real regen (different YAML long-string folding style). Re-ran npm run gen + docs' gen:api and kept that canonical output instead. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): drop the dead access-log email backfill amneziawgEmailIndex keyed peers by "<tag>|<tunnel IP>", a scheme built for the retired TPROXY bridge where the peer's decapsulated tunnel address survived as the access log's from-address. The embedded architecture relays through a loopback SOCKS5 dial, so every AmneziaWG log line's from-address is 127.0.0.1:<ephemeral> and the lookup could never match: the index was rebuilt on every log view just to miss. Remove the index, its GetXrayLogs wiring and its test. If per-line emails are wanted back, the relay would have to publish a local-port->email registry for the viewer to resolve loopback sources. * fix(api): generate AmneziaWGLogs/PeerActivity schemas instead of hand-writing them The amneziawglogs endpoint's response structs were missing from openapigen's StructAllow, so they were silently absent from every generated schema/example, the endpoints.ts entry carried a hand-written response, and AmneziaWGLogModal.tsx duplicated the shapes as local interfaces - the exact drift the allowlist rule exists to prevent. Allowlist both structs with example tags, point the endpoint at the generated schema, import the generated types in the modal, and sync docs/public/openapi.json. * chore(amneziawg): drop the unreferenced quiccapture package Nothing imports internal/amneziawg/quiccapture and no route exposes it; its package doc justifies the code as a port of frontend/src/lib/xray/i1Generators.ts, which does not exist in this repository, and promises an API round-trip that also does not exist. 1,110 lines of unreachable code with misleading provenance claims. Revert this commit to bring the package back when the live-capture I1 feature and its frontend counterpart actually land. * fix(clients): re-run cross-inbound conflict checks on the serialized writer The new client-level checks - cross-inbound AllowedIPs collisions and AmneziaWG forwardedPorts conflicts - read a fresh DB snapshot, decide, and only then enter runSerializedTx, while lockInbound only serializes writers on the SAME inbound. Two concurrent client creates on two different tunnel inbounds both passed the read and both committed, yielding two peers with one address: the exact check-then-claim race 81cfd857 (#6225) closed for AddInbound, which this PR's own checkAmneziawgnetSocksReverseConflict already cites. Keep the pre-tx pass for fail-fast UX and re-validate inside the transaction, where the single writer makes the answer authoritative. The race test drives two goroutines at two inbounds and demands exactly one winner; it fails with committed=2 when the in-tx re-check is removed. * fix(amneziawg): hot-apply depletion disables like mtproto does applyTrafficMutationBatch special-cases MTProto so a quota/expiry depletion cuts the sidecar immediately, but AmneziaWG fell through to runtime AddUser/RemoveUser - explicit no-ops for this protocol - so a depleted peer kept tunneling until the next 10s reconcile tick. Route it through applyLocalAmneziaWG, whose own contract (re-read committed settings, filter depleted clients, push to the interface) is exactly this case; the comment claiming it mirrors applyLocalMtproto is now true for the depletion path too. * fix(amneziawg): persist cleared DNS fields instead of resurrecting defaults PrimaryDNS/SecondaryDNS marshaled with omitempty, so clearing them persisted settings with no key at all - and the frontend re-parses stored settings through a Zod schema whose .default('8.8.8.8') / .default('8.8.4.4') fire on missing keys, silently repopulating the form on every load and re-persisting the defaults on the next save. Blank is a documented, meaningful state (no DNS line in client configs); drop omitempty so a cleared value survives the round-trip. The regression test normalizes a server block with cleared DNS and fails when the keys are dropped. * fix(amneziawg): accept cleared numeric obfuscation/subnet fields in the form AntD InputNumber emits null when cleared, Zod .default() only replaces undefined, and unlike wireguard.ts - whose optionalClearedInt comment documents exactly this failure mode - the AmneziaWG schema declared subnetCidr and jc/jmin/jmax/s1-s4 as bare z.number() defaults. Clearing any of the eight fields made safeParse reject the null and block the save until the user retyped a value. Absorb null into undefined while keeping each field's schema default, so a cleared field refills its documented default and legacy blobs with absent keys behave as before. * fix(amneziawg): guard the third .conf emitter against newline injection The review-round fix added the newline guard to amneziaWGConfigText (Go) and buildAmneziaWGClientConfig, but genAmneziaWGConfig in inbound-link.ts - the third of the three emitters its own comment says must not drift - still rendered privateKey/primaryDns/secondaryDns/remark unescaped, so a newline there injected a config line (e.g. a rogue PostUp) into the inbound form's downloaded .conf. Add the same guard, plus the regression tests the original fix shipped without: all four fields on the Go and both frontend emitters go red if any guard is removed. * test(amneziawg): pin the S3/S4 floors the TS drift guard claims to mirror The test's docstring says it mirrors internal/amneziawg/params_test.go, but it asserted S3>=8/S4>=4 while the Go test and both generators pin 12/12 - the floor ValidateObfuscation enforces whenever a header protection key is set, which this generator always sets. A regression narrowing the TS floors into 8-11/4-11 would have passed the drift guard and produced configs the backend rejects on save. * docs: restore the pia repo-map entry and document the AmneziaWG subsystem Merging main dropped CLAUDE.md's internal/pia/ bullet (added by #6272) while resolving the repo-map conflict - the package itself is untouched. Restore it, add the missing map entries for the two packages this branch introduces (internal/amneziawg/, internal/amneziawgnet/), bump the cron count, and give amneziawg_job its row in architecture.md's 5.4 table. * chore(amneziawg): correct comments stranded by the architecture pivot ae77c7e9's cutover to the embedded gVisor path deleted the kernel-module code but left several comments describing it in the present tense: hot_diff.go cited the removed service.amneziawgEgressStreamSettings and wrongly claimed AmneziaWG is the only sockopt.tproxy source (tunnel's TProxy mode is the live one the guard protects), socks_config.go pointed at the deleted EgressBasePort/EgressPortForInbound, manager.go referred to the deleted Manager and its fingerprinting as live code, web.go's cron registration claimed the job scrapes traffic (its own doc says it does not), and types.go capped ContentPaddingAddition at uint16 when validation and upstream both use uint32. * style(lint): satisfy gofumpt/goimports so make verify is green json_service.go's two 'Tag: "proxy"}' literals came in with main's own cc245a90 formatting commit and fail the repo's gofumpt gate for everyone; the import grouping in inbound_amneziawg.go is from the serialized-writer fix on this branch. --------- Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com> Co-authored-by: Sanaei <ho3ein.sanaei@gmail.com> Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
2026-08-23 17:41:15 -07:00
# resolve_latest_tag prints the latest stable release tag. It prefers the web
# releases/latest redirect, which is not subject to the unauthenticated API's
# 60 req/h-per-IP limit that trips shared CI/CGNAT addresses (the install then
# fails with "Failed to fetch x-ui version"), and falls back to the API.
resolve_latest_tag() {
local url tag
url=$(curl -sSLI -o /dev/null -w '%{url_effective}' --retry 5 --retry-delay 3 --connect-timeout 15 --max-time 60 "https://github.com/MHSanaei/3x-ui/releases/latest" 2>/dev/null)
tag=${url##*/tag/}
if [[ "$tag" != "$url" && -n "$tag" && "$tag" != "latest" ]]; then
echo "$tag"
return 0
fi
curl -Ls --retry 5 --retry-delay 3 --connect-timeout 15 --max-time 60 "https://api.github.com/repos/MHSanaei/3x-ui/releases/latest" | grep '"tag_name":' | sed -E 's/.*"([^"]+)".*/\1/'
}
# Releases publish <asset>.sha256 next to each archive. A mismatch or a failed
# sidecar download aborts the install; only a 404 (releases predating the
# sidecar) is tolerated with a warning.
verify_release_checksum() {
local url="$1" file="$2" sums="$2.sha256" code expected actual
rm -f "${sums}"
code=$(curl -sL --retry 3 --retry-delay 3 --connect-timeout 15 --max-time 60 -o "${sums}" -w '%{http_code}' "${url}.sha256")
if [[ "${code}" == "404" ]]; then
rm -f "${sums}"
echo -e "${yellow}No checksum published for this release, skipping verification${plain}"
return 0
fi
if [[ "${code}" != "200" ]]; then
rm -f "${sums}" "${file}"
echo -e "${red}Failed to download the checksum for $(basename "${file}") (HTTP ${code})${plain}"
exit 1
fi
expected=$(awk 'NR == 1 {print $1}' "${sums}")
actual=$(sha256sum "${file}" | awk '{print $1}')
rm -f "${sums}"
if [[ ! "${expected}" =~ ^[0-9a-f]{64}$ || "${expected}" != "${actual}" ]]; then
rm -f "${file}"
echo -e "${red}Checksum mismatch for $(basename "${file}"): expected ${expected:-<none>}, got ${actual}${plain}"
exit 1
fi
echo -e "${green}Checksum verified: ${actual}${plain}"
}
fix(install): fetch x-ui.sh and unit files from the installed release tag (#6391) * fix(install): fetch x-ui.sh and unit files from the installed release tag install.sh and update.sh pin the panel archive to a release tag but always took x-ui.sh, x-ui.rc and the service units from main, so the management script and the binary of one installation came from different commits: the fail2ban templates and setting flags the script writes drift silently against an older binary, two installs of the same tag differ, and a reviewed or digest-pinned installer still runs unreviewed code from main. Use the same ref as the archive, keeping main only for the rolling dev-latest build. The menu's "update menu" and update_shell paths now fetch the script matching the installed version and fall back to main with a visible notice when no script is published for it. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): fall back to main for files a pinned tag does not publish Review follow-ups. install.sh accepts tags down to v2.3.5, but x-ui.rc only exists from v2.8.4 and the split x-ui.service.* files are newer still, so pinning those to the tag made an Alpine install of an old tag 404 after the previous install was already removed. Probe the tag for each file and fall back to main with a notice when it is missing, as the menu already does for x-ui.sh. The fail2ban auto-setup probe also trusted the exit status of 'x-ui setup-fail2ban', but scripts before v3.4.0 have no such subcommand and exit 0 from the usage banner, so the installer reported a setup that never ran. Skip with a notice when the installed script does not know the subcommand. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): refuse a tag that does not publish a needed script Falling back to main reintroduced the binary/script mismatch the tag pinning exists to remove, and it fired at points where install.sh and update.sh have already stopped and removed the previous installation -- so the quiet path was also the one that could not be undone. Probe the tag instead, before anything is touched, for every file that is always fetched from GitHub (x-ui.sh, plus x-ui.rc on Alpine), and abort with the HTTP status when one is missing. The unit files stay unprobed: they are only fetched when the release tarball omits them, so an old tag that ships x-ui.service inside its tarball still installs. Their existing failure message now names the ref it tried. Also tighten the setup-fail2ban probe to the dispatcher's case arm rather than any mention of the string, which also matches a comment. --------- Co-authored-by: Claude <noreply@anthropic.com>
2026-09-03 07:50:53 -07:00
# Older tags predate some of these files (x-ui.rc arrived in v2.8.4). Serving
# main's copy against an old binary is the mismatch this pinning exists to
# prevent, so probe before anything is stopped or removed and refuse the tag.
require_repo_files() {
local ref="$1" name status
shift
[[ "${ref}" == "main" ]] && return 0
for name in "$@"; do
status=$(curl -sIL --retry 3 --connect-timeout 15 -o /dev/null -w '%{http_code}' "https://raw.githubusercontent.com/MHSanaei/3x-ui/${ref}/${name}")
if [[ "${status}" != "200" ]]; then
echo -e "${red}${name} is not available for ${ref} (HTTP ${status})${plain}"
echo -e "${red}Install a release that ships it, or 'dev' for the rolling build. Your existing installation has not been touched.${plain}"
exit 1
fi
done
}
2026-05-19 04:09:35 -07:00
install_x-ui() {
cd ${xui_folder%/x-ui}/
# Download resources
if [ $# == 0 ]; then
feat(amneziawg): add native AmneziaWG protocol support (#6105) * feat(amneziawg): add native AmneziaWG protocol backend AmneziaWG (WireGuard plus DPI-resistant obfuscation) needs no Docker here — it runs as a genuine kernel interface via awg-quick/awg, managed the same way internal/mtproto manages mtg: one Inbound row is one desired Instance, and a Manager reconciles running interfaces toward the database every 10s (internal/web/job/amneziawg_job.go) plus immediately after a client edit (applyLocalAmneziaWG). Clients reuse model.Client verbatim (the same PrivateKey/PublicKey/ PreSharedKey/AllowedIPs fields WireGuard already uses), so bulk operations, the QR/share-link modal and subscriptions come from the shared inbound infrastructure instead of a parallel implementation. internal/amneziawg owns the obfuscation param generator/validator (ported from coinman-dev/3ax-ui, upgraded to AmneziaWG 2.0's S3/S4 padding and I1 signature packet) and the exec wrapper around awg-quick/awg, with fingerprint-based reconcile (noop / reload-via- syncconf / full restart) mirroring mtproto.Manager so a same-protocol edit doesn't force an unnecessary interface bounce that would drop every peer's connection. Frontend and install.sh's DKMS/awg-tools setup are tracked separately; this is backend-only. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): add frontend support and fix a Go->Zod generator gap Wires the amneziawg protocol through the panel UI the same way every other protocol is registered: a Zod settings schema (nested {server, clients}, matching the Go JSON exactly), the protocol enum, the inbound-form's per-protocol fields component and its tab-visibility allowlist, the default-settings factory, the client schema dispatcher, and the sniffing-capability exclusion (no Xray inbound exists for amneziawg, same as mtproto). Client key/allowedIPs fields are reused rather than duplicated: since AmneziaWG clients are wire-identical to WireGuard clients (same model.Client fields), ClientFormModal renders one shared field block for both, switching only the visible label by which protocol is active. The private-key input also gets a live public-key sync via a new useEffect, because unlike WireGuard's Xray-native inbound (which re-derives its public key at runtime and never stores one), AmneziaWG's server.publicKey is a real persisted field the Go backend reads directly — free-typing a new private key without this would silently save a mismatched keypair. Adds a downloadable per-client .conf (amneziawgConfig.ts, mirroring wireguardConfig.ts) with the obfuscation lines, and an InboundOption.AwgServer field on the Go side so the config builder gets the full server block in one round trip. Along the way, running tools/openapigen surfaced a real bug: it doesn't flatten anonymously-embedded Go structs the way encoding/json does, so ServerSettings embedding Obfuscation20 produced a Zod schema with a nested `obfuscation20` key that never matches the real wire JSON. Fixed by un-embedding (flat fields + an accessor method) and registering internal/amneziawg in the generator's own package list, which had been silently emitting a dangling schema reference. English and Russian translations are complete; the other 10 locale files still fall back to English for the new keys. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): complete frontend parity for the Inbounds list page The Clients page (form, CRUD, QR/config) already worked from the prior commit; this closes the remaining gap on the Inbounds side and in a couple of protocol allowlists that a plain search for existing wireguard/mtproto handling turned up. lib/xray/inbound-link.ts gets amneziawg-specific link/config builders (genAmneziaWGLink/genAmneziaWGConfig, plus the *s fan-out variants) mirroring the wireguard ones — AmneziaWG has no legacy peers-array to fall back to, so these read settings.clients directly and add the obfuscation lines every client must share with the server. Wired into genInboundLinks generically, and into three consumers that call the wireguard builders directly rather than through that dispatcher: QrCodeModal, InboundInfoModal, and InboundsPage's bulk export. ClientInfoModal, ClientBulkAddModal, and the bulk attach/detach modals each had their own protocol allowlist that needed amneziawg added alongside wireguard/mtproto. Two real gaps surfaced by grepping every remaining 'wireguard' / Protocols.WIREGUARD hit in frontend/src rather than trusting the checklist was exhaustive: - useInbounds.ts's TRACKED_PROTOCOLS gates the deactive/depleted/ expiring/online client counts shown per inbound on the list page; without amneziawg those counts would silently read zero. - inbound-tag.ts is an explicit client-side mirror of the Go backend's port_conflict.go (the file says so itself: "Keep in sync"). It still only special-cased wireguard for UDP, so an amneziawg inbound would have fallen through to the TCP default and disagreed with the backend's own port-conflict math. Also finishes translating the AmneziaWG UI strings into the 11 locale files that were still falling back to English (ar-EG, es-ES, fa-IR, id-ID, ja-JP, pt-BR, tr-TR, uk-UA, vi-VN, zh-CN, zh-TW), matching en-US/ru-RU key-for-key (26 new keys, verified by count in every file). Not run anywhere: npm run typecheck / build. This machine has neither Node nor npm, so nothing here has compiled — reviewed by hand plus brace/paren balance checks and cross-referencing the generated Zod/TS types. Treat this as needing a real typecheck before shipping. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * docs(install): note that AmneziaWG kernel module install is still manual Tracked separately (not yet ported into this script) — see coinman-dev/3ax-ui's install_amneziawg for the reference approach (ppa:amnezia/ppa). Also serves as a real, path-filter-matching change to get the previous empty commit's CI trigger to actually fire — release.yml's push trigger is paths-scoped and an empty commit changes no files, so it never matched. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): add a button to randomize obfuscation parameters Mirrors the existing key-regenerate button next to the private key field. Client-side randomization matches the ranges/constraints of GenerateObfuscation20's "default" preset (internal/amneziawg/params.go) closely enough for a form suggestion — the user can still hand-edit any field afterward. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(install): auto-install the AmneziaWG DKMS module + amneziawg-tools Ports install_amneziawg from coinman-dev/3ax-ui's install.sh, adapted to this script's broader distro coverage and NONINTERACTIVE convention: - Ubuntu/Debian/Armbian: ppa:amnezia/ppa (primary, tested path), with a reachability pre-check for the Launchpad PPA host — often blocked by hosting providers, especially Russian VPS — so a flaky network skips the feature instead of hanging apt through several retries. - Fedora/RHEL-family, Arch/Manjaro/Parch: best-effort fallback to plain wireguard-tools (+ AUR amneziawg-dkms via yay/paru when available), with a manual-install pointer. - Everything else: manual-install pointer only. Also installs ndppd and persists IPv4/IPv6 forwarding (for the future IPv6/NDP phase, not yet wired into the panel) and adds a Secure Boot warning at the end of the run, since a DKMS-built module is unsigned and won't load while it's enabled — a common trap on cloud VPS images. Never fatal: the panel installs and runs fine either way, an AmneziaWG inbound just won't bring up its tunnel until the module is present. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): resolve all 3 real CI failures (typecheck/lint/codegen) Found by checking the fork's Actions tab after the last two pushes — the release build passed (it doesn't run these checks) but the separate CI workflow caught three real issues: - golangci-lint (noctx): every internal/amneziawg/manager.go exec.Command call is now exec.CommandContext with a 30s timeout, so a hung awg-quick/awg invocation can't block the reconcile job indefinitely (mirrors internal/mtproto/process.go's own CommandContext usage). - tsc --noEmit: frontend/src/schemas/client.ts's hand-maintained InboundOptionSchema (used by the useClients hook, separate from the auto-generated one in generated/) never got an awgServer field added when the AmneziaWG frontend work was done — every read of inbound.awgServer.* in amneziawgConfig.ts was typing as {}. Added AwgServerOptionSchema, nested (not flattened like wg*) to match what amneziawgConfig.ts already expects. Also guarded server.publicKey in inbound-link.ts's genAmneziaWGLink against the schema's optional type. - codegen staleness: frontend/public/openapi.json is produced by a Node script (gen:api) this machine can't run; hand-applied the exact diff the CI failure log already showed (amneziawg protocol enum entry, ServerSettings schema, InboundOption.awgServer, one example payload), verified as valid JSON. Also confirmed independently by this run: install_amneziawg (previous commit) installed and loaded the DKMS module successfully on both amd64 and arm64 CI runners. The two "Deploy Smoke Tests" failures are unrelated to this change — this fork has only ever published the dev-latest pre-release, and GitHub's /releases/latest API deliberately excludes pre-releases, so the smoke test's no-argument install path (which resolves "latest") has nothing to find. Not a regression; needs an actual tagged release whenever that's wanted. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): Phase 2a — IPv6 support + NDP proxy Adds native dual-stack IPv6 to AmneziaWG inbounds, ported from coinman-dev/3ax-ui's approach: - ServerSettings gets ipv6Enabled/ipv6Subnet/ipv6ExternalInterface; Instance carries the server's own IPv6 address (first host of the subnet) alongside its IPv4 one. - defaultAmneziaWGClients allocates an IPv6 host address per client (second AllowedIPs entry) when the server has IPv6 enabled, reusing allocateWireguardAddress — which needed a real fix along the way: it always suffixed "/32" regardless of address family, which is wrong for an IPv6 host address (needs /128). Now family-aware. - generateServerConfig's PostUp/PostDown gains IPv6 forward-accept rules, proxy_ndp sysctl, and one `ip -6 neigh add/del proxy` entry per enabled peer with an IPv6 address — the lightweight per-client method, not the ndppd-daemon whole-subnet method (not worth the config-file-management complexity at this scale; ndppd itself is still installed by install.sh in case that changes later). - ValidateIPv6Subnet rejects a malformed subnet before save. - Frontend: ipv6Enabled/ipv6Subnet/ipv6ExternalInterface fields on the AmneziaWG inbound form, EN+RU translations, openapi.json/generated/* regenerated (the latter via `go run ./tools/openapigen`, pure Go). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): fill in IPv6 fields missed by the Phase 2a commit Two real gaps the CI caught (both new fields, both my miss): - inbound-defaults.ts's createDefaultAmneziawgInboundSettings() built a server object literal predating ipv6Enabled/ipv6Subnet/ ipv6ExternalInterface — AmneziawgServer's inferred type now requires them (zod .default() fields are non-optional post-parse), so this didn't typecheck at all. - openapi.json's ipv6Enabled property was missing the description the real generator attaches (the Go doc comment covering all three IPv6 fields is attached to the first one) — a one-line diff, but git diff --exit-code doesn't care how small. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): Phase 2b — per-client port-forwarding Admins can now set a per-client ForwardedPorts string (e.g. "80, 443, 8000-8100") that gets DNAT'd + FORWARD'd to that peer's tunnel address via iptables rules in PostUp/PostDown, ported and simplified from coinman-dev/3ax-ui's shared/portfwd. Two decisions worth flagging for future readers: - The iptables --comment tag on each rule is awg-fwd-<fnv32a(email)>, not the raw client email. Email is admin/API-supplied free text that ends up embedded in a shell-executed PostUp/PostDown line; a hash can never carry a shell metacharacter through where raw interpolation could. - The reconcile manager gained a third fingerprint (portFwdFP, next to the existing structural/peers ones). `awg syncconf` only touches the WireGuard peer table — it never re-applies PostUp/PostDown iptables rules — so a port-forward-only change has to force a full awg-quick down+up bounce, same as a structural change, rather than the lighter sync a plain peer add/remove can use. Also fixes a real pre-existing bug found while wiring up IPv6 client allocation in the previous commit's spirit: allocateWireguardAddress always suffixed "/32" regardless of address family, which produced invalid host bits for IPv6 (needs "/128"). ForwardedPorts flows through model.Client -> model.ClientRecord (gorm column wg_forwarded_ports, auto-migrated) -> ToRecord/ToClient/ MergeClientRecord, mirroring the awgServer field's earlier lesson that new fields need checking against a second, hand-maintained persistence-layer struct. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): route a client's traffic through Xray via the Routing page Every enabled AmneziaWG inbound gets its own Xray TPROXY bridge automatically, with no toggle to enable first: a loopback dokodemo-door inbound (sockopt.tproxy) tagged with the AmneziaWG inbound's own real tag, so it's already selectable in the existing Routing page's inbound-tag picker — the same trick the mtproto sidecar's own bridge already relies on (InboundService.GetInboundTags is a plain, protocol-blind SELECT over every inbound row's tag, no dedicated UI plumbing needed). internal/amneziawg's defaultPostUpDown TPROXYs every peer's traffic into that bridge unconditionally; the bridge's port is derived deterministically from the inbound's id (EgressPortForInbound) so the kernel-side reconcile loop and the Xray-config generator never need to negotiate a runtime value between them. injectAmneziawgEgress never generates a routing rule itself — whether a client's traffic goes anywhere beyond Xray's default routing is entirely up to whatever rules the admin adds through the existing Routing UI (pick the AmneziaWG inbound's tag as source, optionally a specific peer's IP via that page's own Source-IP field, and an outbound), exactly the same workflow as routing any other protocol. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): recover orphaned interfaces after an ungraceful exit Two gaps left an AmneziaWG interface stuck outside the manager's control after a crash (kill -9/OOM/panic skips StopAll): - ensureRestart's teardown was gated on the in-memory `exists` map, which is always empty on a fresh process, so a survived interface never got interfaceDown before interfaceUp tried `ip link add` against a name the kernel already had — failing forever and never populating m.ifaces, so traffic accounting silently stopped and the inbound could never be removed. Gate on isInterfaceUp instead, which checks real kernel state rather than this process's own bookkeeping. - An inbound deleted from the database entirely while the panel was down has no entry in `desired` ever again, so it never reaches the per-id cleanup loop in Reconcile (which only walks m.ifaces). Add a one-time sweepOrphansLocked scan of configDir, mirroring mtproto.Manager.sweepOrphansLocked, that tears down and removes any leftover interface/config not in the current desired set. Found by the automated review on MHSanaei/3x-ui#6105 (Finding 1). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * i18n(amneziawg): backfill IPv6/obfuscation/port-forwarding keys in 11 locales Only en-US/ru-RU ever got these 9 keys as each AmneziaWG feature landed (the regenerate-obfuscation button, then Phase 2a's IPv6 fields, then Phase 2b's per-client ForwardedPorts) — the other 11 locale files were never backfilled, so i18next has been silently falling back to English for all of them since Phase 1. Cosmetic-only (never broke anything), but now closed for every shipped locale. * fix(amneziawg): resolve 7 Medium findings from the automated PR review Each is independently reproducible; fixed together since one review pass found all of them. - manager.go: the shared "ip rule add fwmark" policy route had no existence check, so it duplicated in "ip rule show" on every interface bounce (which hostRulesFingerprint forces on any client add/remove/ re-IP). Now checked via "ip rule list | grep -q ..." first. (Finding 2) - params.go: ExternalInterface, IPv6ExternalInterface, and subnetIp/ subnetCidr are interpolated unescaped into a shell-executed PostUp/ PostDown line, but only obfuscation and the IPv6 subnet were validated before save. Added ValidateInterfaceName (a strict charset+length pattern) and ValidateSubnetIPv4 (netip.ParsePrefix), wired into normalizeAmneziaWGSettings. (Finding 3) - amneziawg_job.go: IsAwgInstalled() existed but nothing ever called it, so a host without awg/awg-quick (the Docker image, RHEL, Arch, a failed install.sh PPA step) logged a reconcile failure every 10s forever. Now checked once an inbound actually needs it, warning once instead of spamming. (Finding 4) - client_inbound_apply.go: the WireGuard/AmneziaWG credential carry-forward (added so a metadata-only client edit doesn't rotate keys) never covered ForwardedPorts, so a partial edit -- an API call or Telegram-bot toggle that omits the field -- silently wiped a client's port-forwarding spec. Carried forward and written back the same way the key fields already are. (Finding 5) - manager.go: hostRulesFingerprint keyed each peer on its IPv4 address only, and structuralFingerprint omitted IPv6Enabled/IPv6ExternalInterface entirely, so an IPv6-only change could pick the syncconf reload path (which never re-runs PostUp, leaving a stale NDP-proxy entry) or be a complete no-op. Both fingerprints now cover the IPv6 fields. (Finding 6) - port_conflict.go: the AmneziaWG egress bridge (injectAmneziawgEgress) binds 127.0.0.1:63100+id with no collision check anywhere, since it isn't a database row the ordinary port-conflict query can see -- same blind spot the reserved Xray API port already has its own check for. Added the equivalent check for the AmneziaWG bridge port. (Finding 7) - install.sh: install_amneziawg ran unconditionally for every install/ update, building a DKMS kernel module and enabling host-wide IPv4/IPv6 forwarding whether or not the feature is ever used. Gated behind a new should_install_amneziawg (XUI_INSTALL_AMNEZIAWG=true/false, or an interactive y/N prompt defaulting to no). Also replaced the deprecated apt-key adv with a dedicated keyring + signed-by= on the Debian branch, and guarded its sources.list appends against duplication on a retried install. (Finding 8) Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * feat(amneziawg): make the Xray TPROXY bridge a per-inbound opt-in Addresses Finding 10 from the automated PR review: an always-on TPROXY bridge makes every AmneziaWG tunnel hard-depend on Xray being up (all traffic, including DNS, drops whenever Xray restarts), and forces a full awg-quick down+up bounce on any client add/remove/re-IP, permanently losing the syncconf fast path. Adds ServerSettings.RouteThroughXray (off by default): - defaultPostUpDown only emits the TPROXY/policy-route rules when it's on; a plain AmneziaWG tunnel now has zero Xray dependency out of the box. - structuralFingerprint covers it (toggling it changes whether PostUp/ PostDown contain any TPROXY rules at all -- structural, not a per-peer host-rule). hostRulesFingerprint's IPv4 tracking is now itself conditional on RouteThroughXray (and IPv6 tracking on IPv6Enabled), so an instance that never uses either keeps the syncconf fast path for a plain peer re-IP. - injectAmneziawgEgress only creates a bridge for inbounds that opted in; checkAmneziawgEgressConflict (the Finding-7 fix) now parses each candidate through InstanceFromInbound so a non-routed inbound's port is correctly never treated as reserved. - New inbound-level Switch in the AmneziaWG form; the actual outbound decision is still made entirely through the panel's stock Routing page, same as before -- only whether the bridge exists at all is now a choice. Translation keys added to all 13 locales in the same commit this time, not backfilled later (see Finding 9's lesson). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): resolve 4 Low findings from the automated PR review - manager.go: serverAddress assumed subnetIp always ends in ".0"; a base like "10.8.1.5" was used verbatim as the server's own address, eventually colliding with peer allocation (which starts at .2 upward). Now derives the first host of the actual subnetIp/subnetCidr network via netip, matching serverAddressV6's own approach. A /32 base (no host bits at all) is still used as-is. (Finding 12, partial -- the /16 pool-widening half of this finding only exists on the upstream-pr/amneziawg branch's merged client_wireguard.go, not here; handled separately on that branch.) - manager.go: ensureLocked carried the previous per-peer traffic counters (`last`) forward even through a full restart, but awg-quick down+up resets the kernel's own counters to zero -- the next CollectTraffic computed a large negative delta (clamped to 0), silently discarding real traffic. Extracted the decision into nextTrafficBaseline: only a reload (syncconf) preserves the baseline. (Finding 13) - portfwd.go: exported ForwardedPortsInclude; inbound_amneziawg.go's new checkForwardedPortsConflict uses it to reject, at save time, a client's forwardedPorts that would DNAT the panel's own port or another enabled inbound's port to the tunnel client -- portForwardLines has no destination restriction, so this collision was previously silent. Wired into both the single-client update path and the add-client path (client_inbound_apply.go), plus normalizeAmneziaWGSettings for the whole-inbound save path. (Finding 14) - inbound.go: InboundOption.AwgServer sent the whole ServerSettings struct including PrivateKey to GetInboundOptions callers -- a shared, admin-wide dropdown-filling endpoint the frontend's own AwgServerOptionSchema never reads that field from. Redacted it before assigning. (Finding 11) Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): don't widen the peer address pool past AmneziaWG's own subnet Completes Finding 12 from the automated PR review (the serverAddress half of this finding was already fixed on main and cherry-picked here). This half is specific to this branch: allocateWireguardAddress's /16 pool-widening fallback is an independent addition from upstream's own main that this branch inherited during the cherry-pick rebase -- it doesn't exist on the fork's own main at all, so this fix can't be cherry-picked the normal way and is committed directly here. Widening is safe for WireGuard's own Xray-native inbound (AllowedIPs isn't tied to a strict kernel interface subnet), but AmneziaWG's kernel interface Address is exactly the configured subnet -- an address allocated from the containing /16 once the /24 fills up would be silently unroutable. allocateWireguardAddress now takes an explicit allowWidening bool: WireGuard's own caller passes true (unchanged behavior), AmneziaWG's passes false (fails loudly on exhaustion instead). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * docs(docker): note that AmneziaWG doesn't work in this image Investigated: the image is Alpine-based, and AmneziaWG's own packaging (DKMS module + amneziawg-tools) doesn't target Alpine/musl at all -- unlike the Debian/Ubuntu/Fedora/Arch paths install.sh already handles, there's no package to apk add even with full host network/capabilities. The panel already degrades gracefully (IsAwgInstalled() logs one warning instead of retrying forever), so no code change is needed -- just made the reason explicit at the point where a user would reach for cap_add/ network_mode to try to work around it. * fix(sub): include amneziawg inbounds in subscription links getInboundsBySubId's SQL protocol allowlist never had 'amneziawg' added, so every AmneziaWG client was silently excluded from all three subscription formats (plain/individual links, JSON, Clash) and from the Telegram bot's QR/individual-link buttons, which fetch through the same path. genAmneziaWGLink itself was already fully implemented and already wired into GetLink's dispatch switch -- it just never got a chance to run. Same bug shape as the earlier TRACKED_PROTOCOLS frontend gap: a hardcoded protocol list one entry short. Found while investigating whether the Telegram bot needed AmneziaWG- specific client-management code -- it doesn't (the bot itself is fully protocol-agnostic), but this is the actual root cause of "can't share an AmneziaWG client's config via the bot." Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(inbound): enforce node-eligibility server-side, not just in the UI Investigated multi-node interaction with AmneziaWG: the master's own reconcile (DesiredAmneziaWGInstances) and Xray config generation (injectAmneziawgEgress, the GenXrayInboundConfig protocol skip) all correctly filter on NodeID IS NULL, so a node-assigned AmneziaWG (or MTProto) inbound would never be managed by the master. But nothing stopped one from being created that way: NODE_ELIGIBLE_PROTOCOLS (frontend/src/pages/inbounds/form/InboundFormModal.tsx) only hides the node picker client-side -- a direct API call could set nodeId on an AmneziaWG inbound, which every node then reconciles as an ordinary local inbound (nodes run the identical binary, full cron suite included), leaving it running unmanaged and untracked by the master's own AmneziaWG bookkeeping. Added isNodeEligibleProtocol (inbound_protocol.go), mirroring the frontend's allowlist, and enforced it in both AddInbound (the actually exploitable path -- nodeId comes straight from the request) and UpdateInbound (defense in depth; NodeID is already restored from the stored row there before this check, so it mainly guards against a protocol change on an existing node-hosted inbound). Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): allow TPROXY-marked traffic through a default-deny INPUT chain TPROXY never rewrites a packet's own destination address, only the routing decision. A default-deny firewall whose INPUT chain sanity-checks "is this destination actually local" (UFW's ufw-not-local, via addrtype --dst-type LOCAL, is a concrete example) silently drops the redirected packet before Xray's socket ever sees it -- RouteThroughXray looked fully configured (TPROXY rule present and counting, Xray listening with IP_TRANSPARENT set) yet every peer's traffic vanished with no trace on either side. Adds an idempotent, never-torn-down "iptables -I INPUT 1 -m mark --mark <fwmark> -j ACCEPT" alongside the existing shared policy route, so this works regardless of which firewall manager owns the rest of the INPUT chain. * fix(frontend): give AmneziaWG the same UDP tag and its own tag color The Inbounds list only special-cased isWireguard/isHysteria for the "UDP" network badge, so an AmneziaWG row showed just the bare protocol tag with no transport badge next to it. Added the missing isAmneziawg flag (mirrors isWireguard exactly) and wired it into the same branch. Client-row protocol-color maps in ClientsPage/HostList had no amneziawg entry, silently falling back to grey -- ClientInfoModal already had amneziawg: 'yellow' from earlier work, these two just never got it. * feat(logs): show which AmneziaWG client an access-log line belongs to The dokodemo-door TPROXY bridge every AmneziaWG peer's traffic is routed through has no per-user identity, so Xray's own access log never carries an "email:" token for these lines -- the Access Logs modal showed a blank Email column for every in-*-udp row, even though every other protocol's rows show the client normally. The peer's decapsulated tunnel IP does survive as the log's "from" address, and that IP deterministically maps to exactly one configured peer. Builds a "<inbound tag>|<ip>" -> email index from the same AmneziaWG inbounds already parsed elsewhere (amneziawg.InstanceFromInbound), and fills in Email from it whenever the raw log line didn't have one. * fix(amneziawg): enable sniffing on the TPROXY bridge Domain-based Routing rules could never match RouteThroughXray traffic: an AmneziaWG peer resolves DNS itself, through the tunnel, before ever sending a packet, so the decapsulated traffic TPROXY hands to the bridge is already a bare destination IP with no domain name attached at the network layer. Every other inbound recovers this via sniffing (confirmed working for the stock wireguard inbound, which does have it configured); the bridge never got a sniffing block at all, so only tag/IP/network-based rules could ever match it -- any domain rule above it in the list was silently unreachable. * docs: add an AmneziaWG config page and list it as a supported protocol Closes the PR checklist gap: the feature shipped with zero mention on the docs site. Mirrors reality.mdx's structure (key settings, setup steps, config excerpt) and notes the Docker/multi-node/Telegram-bot caveats the PR itself is honest about not having confirmed. * fix: address the fresh review round on PR #6105 (8 findings) 1. hostRulesFingerprint didn't account for ForwardedPorts when RouteThroughXray was off, so re-IPing a peer with port-forwarding configured left stale DNAT rules pointing at an address the next peer could be handed. 2. Server/client config values (keys, email, I1) were never validated for control characters before being written into the generated .conf; a newline could smuggle a PostUp hook into awg-quick's parser. Added ValidateConfigValue at save time and a sanitizeConfigValue backstop at render time. 3. checkForwardedPortsConflict didn't scope to node_id IS NULL, so a port used only on a different node produced a false collision; also hoisted the panel-port/inbounds lookup out of the per-client loop (portConflictContext) so N clients cost one query, not N. 4. PostDown commands were ";"-joined and abort on the first failure; appendOrTrue makes teardown best-effort so an external firewall flush can't leave DNAT rules to accumulate across bounces. 5. The "ip rule list | grep -q" existence check could SIGPIPE under pipefail and re-add a duplicate rule; switched to grep -c >/dev/null. 6. Ported the vpn:// share-link format (base64url of the plain .conf text, matching the real AmneziaVPN app) onto this branch -- it had only ever landed on our own fork's main, so this PR branch was still on the old amneziawg://+query-params scheme our own docs no longer described. Also corrected the docs' install.sh claim (opt-in/ interactive, not automatic) and stale pre-opt-in comments in route_egress.go. 7. install.sh: Arch's ndppd install used pacman -Syu (full system upgrade) instead of -Sy like every other call in the script; and should_install_amneziawg re-prompted on every `x-ui update` even when awg was already installed. 8. CollectTraffic could clobber a concurrent restart's freshly-reset (empty) traffic baseline with stale pre-restart counters, since getPeerStats runs lock-free; now checks pointer identity before writing back. sweepOrphansLocked permanently disabled itself on a transient os.ReadDir failure instead of allowing a retry. go build/vet/test and frontend typecheck/lint/build/vitest all pass. * fix(install.sh): check the live sysctl value, not sysctl.conf text Reviewer feedback (cherts, PR #6105): grepping /etc/sysctl.conf for the setting name is unreliable -- many distros split sysctl config across /etc/sysctl.d/*.conf, and /etc/sysctl.conf can be a symlink into that directory, so the check can miss an already-active setting (harmless duplicate append) or match a disabled/commented line (forwarding silently stays off). Query the live value via `sysctl -n` instead, which is accurate regardless of which file set it. Applied the same fix to both the IPv6 and IPv4 checks for consistency. * fix: update inbound_amneziawg.go to the split buildInboundForLocalRuntime Same fork-only-file blind spot as the one caught on our own main after the 3.6.0 sync: upstream split buildRuntimeInboundForAPI into buildInboundForNodePush / buildInboundForLocalRuntime (part of the node-sync client-deletion fix, 5bc81dfd), updating every call site it could see. This file doesn't exist upstream, so it kept calling the old name even after the branch merged in that commit. * fix(frontend): recognize AmneziaWG's vpn:// scheme in share-link labels The shared link-tag/label helper (used by the client info modal, QR modal, and subscription page) had no entry for the vpn:// scheme AmneziaWG links use, so it fell through to the generic fallback: a plain "Vpn" tag with no color, and an empty remark/port that made the row's title fall back to "Link N" instead of the inbound's actual name:port — unlike every other protocol, which shows its real tag and label. vpn:// links are base64url of a plain .conf text (matching the real AmneziaVPN app's own share-link format), not a structured URL, so there's no query string or #hash to read a remark/port from. Decode the payload and pull the remark/endpoint back out of the .conf text directly instead. * fix(xray): force a full restart for TPROXY inbounds, never hot-add them Real incident: an AmneziaWG inbound with RouteThroughXray enabled lost all internet on that connection after a migration. Root-caused on the live box -- iptables TPROXY counters were incrementing (packets correctly redirected to 127.0.0.1:63110), but nothing was actually listening there (ss showed nothing on that port) until a full `systemctl restart x-ui`, after which the bridge came up immediately. Xray-core's gRPC AddInbound reports success for a new sockopt.tproxy inbound (internal/amneziawg's own Xray egress bridge is the only kind this fork ever generates) but doesn't reliably bind a working listener for it outside of process startup -- the bridge silently never comes up, and RouteThroughXray traffic goes nowhere until the next full restart happens to occur for an unrelated reason. diffInbounds already has this exact defensive pattern for REALITY inbounds ("a gRPC remove+add does not reliably rebuild the REALITY authenticator"), just never extended to TPROXY, and only in the already-existing-then-changed branch -- the "brand new inbound" branch had no such guard at all, which is exactly the path a freshly-enabled RouteThroughXray bridge takes. Added inboundUsesTproxy and wired it into both branches. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): flag Xray for resync when a peer edit changes qualifying state Real production bug, root-caused on iiadmin-vps: updateAmneziaWGInbound/ AddInbound/DelInbound only ever updated the kernel interface via amneziawg.GetManager() -- they never called SetNeedRestart the way every other protocol's mutation path does (client_crud.go, inbound.go, etc. all do). injectAmneziawgEgress's TPROXY bridge inbound depends on InstanceFromInbound finding at least one qualifying peer plus RouteThroughXray, so an edit that flips that (first peer added, last one removed, RouteThroughXray toggled on) previously required a full panel restart before the bridge actually got created, with no error anywhere: the kernel interface would handshake fine, but traffic redirected into the bridge's TPROXY port went nowhere because nothing was listening there. diffInbounds/inboundUsesTproxy already correctly force a full restart for a brand new TPROXY inbound (bdee0a20) -- that part was never the bug. The gap was entirely upstream: nothing ever told Xray a resync was even needed. * fix(clients): reject AllowedIPs already used on another WireGuard/AmneziaWG inbound defaultWireguardClients/defaultAmneziaWGClients only ever checked uniqueness against their own inbound's client list, so two inbounds sharing a subnet (same protocol or not) could silently hand out or accept the same address -- the exact scenario behind a real duplicate-IP incident where a WireGuard and an AmneziaWG client both ended up on the same address. otherTunnelAllowedIPs now collects every address already claimed on every other tunnel inbound and folds it into both the auto-allocation pool and the manual-entry collision check, naming the other inbound in the error when it fires. * fix(frontend): add the missing AmneziaWG config download on the sub page The subscription page already gave WireGuard links their own "Config" block (copy/download/QR of the actual .conf, via wireguardConfigFromLink reversing the wireguard:// query params) but had no equivalent for AmneziaWG's vpn:// links -- its isWireguardLink gate never matched them, and no reverse-parse helper existed for this page specifically. Every other surface (InboundInfoModal, ClientInfoModal, ClientQrModal) already had this parity; this was the one page that didn't. Fixed by adding amneziawgConfigFromLink (inbound-link.ts), simpler than its WireGuard counterpart since a vpn:// payload already *is* the plain .conf text -- just base64url-decode it, no query-param reconstruction needed -- and wiring it into SubPage.tsx alongside the existing WireGuard block, reusing the same pages.clients.amneziaWgConfig label the other three surfaces already use. * fix(xray): force a full restart for password-auth SOCKS5 hot-apply Real production incident: editing a client under an AmneziaWG inbound left its embedded SOCKS5 relay's settings byte-different (a new account list), and Xray's gRPC remove+add hot swap silently dropped the account for a peer whose email contained non-ASCII characters -- its tunnel kept handshaking fine but all its traffic got rejected at the SOCKS5 layer, while every other peer on the same relay was unaffected. A full restart (reading the same JSON straight from disk) always produced the correct account list. socks isn't in userDiffableProtocols (that only covers vless/vmess/trojan's clients+email shape, not accounts+user), so any settings drift on this inbound fell through to the generic remove+add path. Forces a restart instead, the same defensive choice already made for REALITY and TPROXY -- scoped to auth:"password" specifically so the other, noauth SOCKS5 bridges (panel/node/mtproto egress) keep the cheaper hot path. * Fix Attach reusing one identity's address across wg/awg inbounds ClientService.Attach deliberately copies one identity's stored AllowedIPs into every WireGuard/AmneziaWG inbound it's attached to in the same call, so the same person gets the same tunnel address on every protocol they use. Its loop calls addInboundClient once per inbound, and each of those independently computes otherTunnelAllowedIPs -- so by the second inbound in the batch, the first inbound's just-written copy of this identity's own address looked like a cross-inbound collision against itself. Real production symptom this caused: detaching then re-attaching a client to both wg and awg failed with "wireguard: allowedIPs entry X is already used by a client on inbound 'awg' (#N)" -- the exact address the identity is supposed to keep, rejected as if it belonged to someone else. Add a selfEmails exclusion to otherTunnelAllowedIPs and populate it from the client(s) being processed at the one real call site. Safe unconditionally: ClientRecord.Email is globally unique, so a match can only ever be this same identity's own entry on a sibling inbound, never a genuine different client's address. Reproduced the underlying mechanism live (manual entry correctly rejected as a cross-inbound collision; fresh auto-allocation correctly avoided a used address) before writing the fix, to confirm the guard itself works and the bug is specifically in how Attach's per-inbound calls interact with it. * Attach: allocate fresh when re-attaching with no active tunnel The previous fix (82cc69f5) made Attach's own address-reuse correctly not collide with itself across inbounds -- but it still always reused an identity's stored AllowedIPs verbatim, even when that identity currently has zero WireGuard/AmneziaWG attachments at all. A real report from testing this live: an identity fully detached from both its wg and awg inbounds, then re-attached, got its old address back even though several lower addresses were free -- because nothing about being fully detached ever cleared the stored value Attach copies from. Add hasTunnelAttachment, checked once against the identity's CURRENT inbound set before Attach's loop runs: if none of its current inbounds is WireGuard/AmneziaWG, clear the stored AllowedIPs so this attach allocates fresh (matching what a brand-new client would get) instead of resurrecting an address nothing reserves anymore. Left alone when the identity already has an active tunnel elsewhere, so extending it to a second protocol still keeps a consistent address. * Fix TestOtherTunnelAllowedIPsExcludesSelfEmail's own test setup CI caught this: the "genuinely different client" (other@wg) was seeded onto the SAME inbound passed as excludeID, which otherTunnelAllowedIPs already excludes entirely regardless of the selfEmails fix -- so the assertion that its address is still reported could never have passed, proving nothing either way. Move it onto the sibling inbound alongside shared@id, which is what the test actually needs to exercise (two clients on one sibling, one excluded by email, one not). * Attach: never inherit an address that doesn't fit the target inbound hasTunnelAttachment (from the earlier fix, commit 51067f16) only asked "does this identity have ANY tunnel attachment", treating that as license to reuse its stored address verbatim on every inbound being attached. Real production case this missed: an identity's stored address came from WireGuard's own fallback subnet (10.0.0.0/24, used when that inbound has no other clients to infer a base from), then got attached to a second, AmneziaWG inbound configured for a completely different subnet (10.8.1.0/24). defaultAmneziaWGClients's already-set-AllowedIPs branch only checks for collisions, never subnet membership, so the mismatched address was accepted silently -- producing a peer that can never actually connect, since an AmneziaWG address must fall inside the kernel interface's own configured subnet to be routable at all. Add addressesFitAmneziaWGInbound, checked per inbound inside Attach's loop: if the inherited address doesn't fit the SPECIFIC inbound being attached, clear it just for that one so it gets a fresh, valid allocation instead, while other already-attached inbounds keep their existing values. WireGuard has no equivalent strict subnet requirement (allocateWireguardAddress can widen to a fallback pool for it), so this only ever constrains AmneziaWG targets. * Give WireGuard an explicit, admin-configurable subnet field WireGuard previously had no configurable subnet at all -- only an implicit one, either inferred from existing clients' own addresses (wireguardAllocationBase) or a hardcoded 10.0.0.0/24 fallback when none exist yet. AmneziaWG, by contrast, has always had a real server.subnetIp/subnetCidr field in its settings, editable in the UI. User request: give WireGuard the same treatment. Backend: explicitWireguardSubnetBase reads an optional subnetIp/ subnetCidr pair from the inbound's own settings JSON (mirroring AmneziaWG's defaultAmneziaWGSubnetBases). defaultWireguardClients checks it first; only when unset does it fall back to today's inference-from-existing-clients behavior, so an inbound saved before this field existed keeps working exactly as it always has. Frontend: subnetIp/subnetCidr added to WireguardInboundSettingsSchema and the inbound form (mirroring AmneziaWG's own field layout/labels), with a real default (10.0.0.0/24, the same value the backend already fell back to) seeded for newly created inbounds so the field starts populated and editable rather than blank. Translated across all 13 locales. This also structurally closes the class of bug fixed in 82cc69f5/291c47b3: with wg and awg subnets explicit and independently controllable, an admin who wants matching addresses across both protocols can configure them to actually agree, instead of one silently inheriting the other's incompatible range. * Split the client edit form's AllowedIPs into per-protocol fields A client attached to both WireGuard and AmneziaWG shared one AllowedIPs form field with a dynamically-switching label, so its two genuinely different addresses could never both be shown or edited correctly. Worse, Update/Create broadcast that one shared value to every attached wg/awg inbound with no subnet-fit check, so an ordinary edit save could silently overwrite one protocol's address with the other's -- the same bug class already fixed for Attach, but reachable from any client edit. model.Client gains an optional AllowedIPsByInbound map so a caller can send distinct values per inbound; Update/Create honor it and, when it's absent, clear a shared value that doesn't fit an AmneziaWG inbound's own subnet instead of writing it through. A new TunnelAllowedIPsByInbound read path feeds the real per-inbound address to the client edit form via GET, which now renders two separate, correctly-labeled fields whenever both protocols are attached (unchanged single dynamic field otherwise). * Regenerate openapi.json for the new allowedIPsByInbound field Follow-up to 878ee839: gen:zod (frontend/src/generated) was already regenerated and committed, but gen:api (frontend/public/openapi.json) wasn't, so CI's codegen drift check failed. * Fix build breakage from merging upstream main: Update() gained a limitHwid param Two of our own AllowedIPs tests (not present upstream, so the merge never flagged them as conflicting) still called the old 3-arg Update(inboundSvc, id, client) -- upstream's hardware-ID-limit feature added a required limitHwid parameter that every other caller in this package already passes. Also drop createDefaultInboundSettings from InboundsPage.tsx: the merge conflict resolution kept the import, but upstream's clone-payload refactor (buildClonePayload, inbound-clone.ts) already calls it internally now -- this file doesn't need it directly anymore. * Fix real bug: AmneziaWG clients rejected as "empty client ID" in 3 places Three switch statements on inbound.Protocol handle "wireguard" explicitly (checking client.PublicKey) but fall through to the default case for "amneziawg" (checking client.ID, which AmneziaWG clients never set -- they use PublicKey/Email like WireGuard, not the VMess/VLESS UUID field). This is what the 4 AllowedIPs tests were actually catching: UpdateInboundClient's newClientId derivation hit this same default branch, so every Update() on an AmneziaWG client returned "empty client ID" before ever reaching the AllowedIPs logic being tested. Fixed by adding "amneziawg" alongside "wireguard" in each switch: addInboundClient's per-client validation, UpdateInboundClient's newClientId derivation, and AddInbound's per-client validation (the third one wasn't hit by these tests, but has the identical bug -- creating a brand-new AmneziaWG inbound with a client attached would fail the same way). * refactor(amneziawg): rename Obfuscation20 to Obfuscation31, drop the dead mobile preset Mechanical rename ahead of the AmneziaWG 3.1 parameter work: the type, generator and prose all said 2.0, and the "mobile" generator preset was reachable only from its own test. No behavior change. * feat(amneziawg): AmneziaWG 3.1 obfuscation parameters (backend + generated schemas) Adds the 3.1 parameter surface to the inbound settings and both Go config emitters: I2-I5 signature packets, HeaderProtectionKey (base64 32-byte, shared server<->client), ContentPaddingAddition, the five handshake-timing randomization ranges (RekeyAfterTime/RekeyTimeout/RejectAfterTime/ KeepaliveTimeout/MaxHandshakeAttempts), and the RandomTrailers/ DisableCookies switches. Freshly generated sets fill everything except I2-I5 (matching Amnezia's own generator) with jittered ranges bracketing WireGuard's stock timing constants; every reject window starts >= 30s above the rekey window by construction. Empty fields stay off the wire, so blanking a field disables just that feature. Validation generalizes the H1-H4 range checker for the new uint32-range fields, requires min 1 on timers, cross-checks rekey-vs-reject, and demands a real 32-byte base64 header-protection key. The manager warns once per process when the installed awg tools predate 3.1 but an inbound uses 3.1 parameters (awg-quick rejects unknown keys with a generic error otherwise); apply still proceeds. Requires amneziawg-tools v3.1.20260812+ / module or amneziawg-go v3.1.20260814+ on the host. * feat(amneziawg): emit and randomize 3.1 parameters in the frontend Both client-config emitters (the vpn:// link builder and the clients-page .conf builder) now carry the 3.1 [Interface] lines in the same order as the Go emitters. The obfuscation randomizer moves out of InboundFormModal into a shared lib/xray/amneziawg-obfuscation.ts that also fills the new fields, and createDefaultAmneziawgInboundSettings switches from static values to that generator — a fresh inbound now really gets the unique fingerprint the docs promise instead of the same jc=5/jmin=10 set on every install. Schema parse-time defaults for the new fields stay ''/false on purpose: real values come only from the generator, so resaving an inbound never mutates its stored parameters. A new parity test pins the hand-written AmneziawgServerSchema to the generated ServerSettings key set, so a field added on one side can no longer silently vanish from configs. * feat(amneziawg): 3.1 form fields and translations Inbound form gains inputs for I2-I5, HeaderProtectionKey (filled by the existing obfuscation Regenerate button), ContentPaddingAddition, the five timing ranges, and the RandomTrailers/DisableCookies switches; the MTU input picks up the min=1 its schema already enforced. All 13 locales get the 19 new keys and drop the "2.0" branding from the s3/s4/i1 labels. * docs(amneziawg): document 3.1 parameters; install.sh kernel/version notes The AmneziaWG page's obfuscation section moves from the 2.0 to the 3.1 parameter set: table rows for I2-I5, HeaderProtectionKey, ContentPaddingAddition, the timing-randomization ranges and the RandomTrailers/DisableCookies switches, a requirements callout (tools v3.1.20260812+, module/awg-go v3.1.20260814+, Linux 6.7+ for the DKMS path), and a sample client .conf that matches what the panel actually emits (including the DNS defaults and PersistentKeepalive it always had). install.sh warns before a DKMS build on a pre-6.7 kernel and after any install that left pre-3.1 amneziawg-tools on PATH. Also updates the hosts API operation paths ({id} -> {groupId}) in the stale ru/zh/fa reference pages: syncing docs/public/openapi.json for the new AmneziaWG schema fields surfaced that rename, which had never been copied over, and the docs build fails on paths missing from the spec. * fix(amneziawg): reject control characters and canonicalize 3.1 range values Adversarial review of the 3.1 work surfaced a validation gap: base64.DecodeString silently ignores CR/LF, so a header-protection key that picked up a line wrap in transit decoded to a valid 32 bytes, passed validation, and was emitted verbatim into every client config — where the orphan second line breaks the import while the server (whose emitter strips control chars) keeps running with the correct key. The key and range validators now reject control characters outright. Also from the same review: range values are canonicalized on save ("110 - 140" -> "110-140", whitespace-only collapses to feature-off, closing a case where the server conf rendered an invalid blank-value line the client emitters omitted); the rekey/reject invariant is now enforced against WireGuard's 120s/180s defaults when only one side is set; and the structural fingerprint joins on "\n" instead of "|", which is a legal I1-I5 character and made adjacent free-text fields join-ambiguous. * fix(install): resolve latest release tag via web redirect to dodge API rate limits The non-interactive install smoke test resolved the release version through the unauthenticated GitHub API (api.github.com/.../releases/latest), which allows only 60 requests/hour per IP. The test installs twice in one run, and on shared CI runner IPs the second call gets rate-limited, returns no tag_name, and install.sh treats an empty version as fatal (exit 1) — the same "Failed to fetch x-ui version" real users hit behind CGNAT/shared addresses. resolve_latest_tag() now reads the tag from the github.com releases/latest web redirect (not subject to the API rate limit), falling back to the API only if the redirect yields nothing. Verified with the real deploy/test/smoke-noninteractive.sh (two installs, both green). * fix(amneziawg): three review findings on #6105, plus a comment trim 1. A peer's allowedIPs reached the generated .conf unvalidated and unsanitized, unlike email/publicKey/preSharedKey which normalizeAmneziaWGSettings already guards. A newline in an entry let a following "[Interface]" re-open the interface section, whose "PostUp = ..." awg-quick then runs as root on the next apply. Reproduced end to end against generateServerConfig. The save path now rejects and canonicalizes through normalizeWireguardAllowedIPs, and the render path sanitizes as a backstop for rows predating the validation (an upgrade, a restored backup, a direct DB edit). H1-H4 get the same render-time sanitize, and the two NIC name fields a plausibility check, since stripping control characters alone would still let a shell metacharacter into a root-executed PostUp line. 2. EgressPortForInbound is 63100 + inbound id, so an id past 2435 derives a port above 65535 -- and Xray rejects the whole generated config over one invalid port, taking every other protocol down with it. It now reports ok=false past the range, and both the Xray bridge and its TPROXY rules are skipped instead of emitting an impossible port. 3. The downloadable AmneziaWG .conf read ClientRecord.allowedIPs, a single shared column that holds the WireGuard address for an identity attached to both protocols -- the exact ambiguity tunnelAllowedIPs was added to resolve for the edit form. The info and QR modals already hydrate that field, so they now pass this inbound's own address to the builder. Also trims the comment blocks in the files touched here to the 2-line guidance in CLAUDE.md: internal/amneziawg alone carried 423 comment lines in over-long blocks against 118 for the comparable internal/mtproto, and is now at 110. Every non-obvious constraint is kept (the kernel S1/S2 rule, why PostDown is best-effort, why grep -c and not -q, why the fingerprints split three ways); the narration is gone. Two hot_diff.go comments pointed at an internal/amneziawgnet package and an injectAmneziawgnetSocks function that exist nowhere in the tree; the checks themselves are unchanged. * feat(logs): add an AmneziaWG log view to the overview The overview has an access-log view for Xray but nothing for AmneziaWG, so when a tunnel misbehaves there is no way to see it from the panel at all. A kernel tunnel logs no per-request lines, so the equivalent view is built from the two things it does expose: - Live per-peer activity from `awg show <iface> dump`, joined to the client email through the desired peer set: last handshake, endpoint, allowed IPs, cumulative transfer and online state, newest handshake first. - The panel's own AmneziaWG event lines (interface up/down, awg-quick failures, the pre-3.1 tools warning), which are what actually explain a peer being absent from the table. POST /panel/api/server/amneziawglogs/:count serves both, with the same count + filter contract GetXrayLogs uses, and the modal mirrors XrayLogModal's toolbar, auto-update, mobile cards and download. The action-bar button is gated on a new status.amneziawg.configured, which stays true while an inbound exists but its interface is down -- exactly when the event lines matter. Verified against a running panel: the endpoint returns the peer table and real event lines ("awg/awg-quick not found on PATH", "create config dir: permission denied"), and count and filter both narrow as documented. One of those lines surfaced a Debugf that had been rendering as "for inbound1:amneziawg:"; fixed here since it is now user-visible. * fix(amneziawg): stop double-counting a routed inbound's traffic injectAmneziawgEgress tags its Xray bridge with the AmneziaWG inbound's own tag, so the stock Routing page can target it. Xray therefore reports that bridge's bytes under the inbound's tag, and XrayTrafficJob feeds them to AddTraffic -- which accumulates -- on top of the same bytes AmneziaWGJob already reported from `awg show dump`. An inbound with routeThroughXray on counted roughly twice its real traffic, which also inflates the quota checks that read the same counters. The awg counters are the complete measure: every peer, whether or not TPROXY routed it, and the same wire bytes the per-client totals are built from, so they stay and the Xray rows are dropped. Per-client stats were never affected -- a dokodemo-door bridge has no per-user identity, so Xray emits no user>>>email rows for it. Filtering happens before every consumer, so the DB totals, the external traffic inform and the dashboard's live speed all read one source per inbound. The set of bridge tags now comes from a predicate shared with injectAmneziawgEgress itself, with a test that pins the two together -- naming one tag too few doubles the traffic again, one too many makes real traffic vanish. * fix(amneziawg): align the three .conf emitters on one peer field order The panel builds an AmneziaWG client .conf in three independent places, and they disagreed: buildAmneziaWGClientConfig put PresharedKey right after PublicKey (wg-quick(8)'s own order, and what both WireGuard emitters on the clients side already use), while genAmneziaWGConfig and the Go amneziaWGConfigText put it after Endpoint. A user comparing a subscription link against a downloaded .conf sees the difference immediately, and the generators are exactly the kind of parallel implementation CLAUDE.md warns about drifting. Moves the two outliers onto the wg-quick order. Also drops the stray trailing newline that only appeared when PersistentKeepalive was set, so a config now always ends on its last set field whichever that is -- the same shape all three emitters produce for the same client. Parsing is unaffected either way (the format is order-insensitive, and the AmneziaVPN app reads it as a flat key-value bag), so this changes only the rendered text. Adds a test on each side that pins the peer block's field order, since nothing previously asserted it. * refactor(amneziawg): switch to the embedded amneziawg-go/gVisor architecture Replaces the kernel-module (DKMS) + awg-quick + TPROXY backend with the fork's own embedded design: amneziawg-go runs in-process over a userspace gVisor netstack, and each peer's decapsulated traffic relays into its own loopback Xray SOCKS5 inbound, so Xray's native stats/sniffing/routing work for free instead of through hand-rolled bridges. No kernel module, no DKMS, no Secure Boot conflicts, works the same in a container as on bare metal. - internal/amneziawgnet: new package (Device/UAPI, gVisor netstack, TCP/UDP forwarding, SOCKS5 relay, peer identity, IPv6 host-alias egress identity, per-client port-forwarding) - amneziawg-go v3.1.20260814 + gvisor. - internal/amneziawg: keep the reusable protocol-shape types/validation (Instance/Peer/Obfuscation, InstanceFromInbound); drop the OS-shellout half (awg-quick, TPROXY policy routing, NDP proxy, peer-stats parsing). - internal/web/service: rewire the 5 integration points (job, runtime, client-apply, web shutdown, xray config) from the old manager to the new one; the AmneziaWG log view is rebuilt on the embedded Device's own UAPI dump (extended to carry endpoint/AllowedIPs) instead of `awg show dump`. - install.sh: drop DKMS/ndppd/TPROXY/Secure-Boot installer code (~250 lines) - an entire recurring class of installer fragility goes away. - frontend: drop the now-meaningless routeThroughXray toggle (the relay is always on); keep the field in the Zod schema, unexposed, so it isn't silently stripped from stored settings on next save - two regression tests deliberately depend on the Go struct still carrying it. - docs/i18n: rewrite amneziawg.mdx for the new architecture; drop the dead routeThroughXray translation keys across all 13 locales. Real production throughput (embedded core datapath, isolated bench, same box the kernel-module path was measured on): ~296 Mbit/s up, ~640 Mbit/s down, vs. 414.69 MB/s (~3.3 Gbit/s) for the kernel module on the same hardware - a real gap, tempered by this being single-stream/no-SOCKS5-hop and most VPN traffic being latency-bound rather than throughput-saturating. * fix(amneziawg): restore the branch's own Obfuscation31 shape + 2 CodeQL findings The previous push's wholesale-copy of types.go/params.go from the fork's main branch pulled in that branch's own independent (and incompatible) naming for the same AWG 3.1 feature set: Obfuscation20/GenerateObfuscation20 instead of this branch's already-shipped Obfuscation31/GenerateObfuscation31, and a missing CanonicalizeUintRange -- broke every Go CI job (the whole matrix fails to compile when any one package doesn't, which is why govulncheck/ golangci/postgres-durable-first/race all failed identically, not just go-test). Restores params.go/params_test.go verbatim from this branch's own last commit (a strict superset of validation: it already cross-checks rekey vs. reject timing windows, which the copied version never did) and folds the 3.0/3.1 fields (HeaderProtectionKey, ContentPaddingAddition, the 5 timing fields, RandomTrailers/DisableCookies) into Obfuscation31 itself, matching the original struct exactly instead of as separate top-level Instance fields. instance.go, the two amneziawgnet call sites, and 7 amneziawgnet test files updated to match. Also drops the one test (sanitizeConfigValue) that only ever served the retired kernel-module .conf writer -- correctly not ported, so the test testing it shouldn't have been copied either. Also fixes 2 CodeQL findings the same push surfaced: a clamped uint64->int64 conversion for the new log view's live byte counters (server.go), and an unneeded len+len sum feeding a slice pre-size in the v6-egress outbound merge (xray.go) -- append already grows correctly without it. * chore(amneziawg): regenerate frontend schemas for updated doc comments npm run gen was missed after the previous commit's types.go doc-comment edits (Obfuscation20 -> Obfuscation31, ValidateHeaderProtection -> ValidateObfuscation in the prose) -- openapigen bakes those comments into the generated schema's description field, so the committed frontend/src/generated/schemas.ts and openapi.json still had the old wording. codegen's git-diff-exit-code check caught it correctly. * fix(amneziawg): narrow 2 test fixtures that collided with MaxForwardedPorts TestCheckForwardedPortsConflict_CollidesWithEnabledInboundPort and ..._NoCollisionWhenPortsDontOverlap used "8000-8100"/"9000-9100" as their ForwardedPorts fixture -- 101 ports each, one over MaxForwardedPorts (100). The cap check (checkForwardedPortsConflict, added this session alongside the SOCKS-phantom-port check) fires first, so both tests got "more than 100 forwarded ports" instead of ever reaching the collision logic they're actually testing. The cap itself has its own dedicated boundary test already; these two just needed a narrower range that still covers/misses port 8080 as intended -- 8075-8085 and 9075-9085, 11 ports each. * fix(amneziawg): checkAmneziawgnetSocksConflict had no receiver in its new home My merge-conflict resolution kept this as a method call (s.checkAmneziawgnetSocksConflict) inside checkPortConflictTx, a plain function with no *InboundService receiver -- upstream's #6225 fix moved the port-conflict check out of the (s *InboundService) method and into this new tx-scoped free function, and I didn't notice the call site needed to change shape too. CI caught it immediately (undefined: s); nothing in this specific package can be locally verified past internal/database's own unrelated, pre-existing CGO build issue on this dev machine. Since the signature had to change either way, folded in the fix already flagged as a separate follow-up: checkAmneziawgnetSocksConflict now takes the caller's db handle instead of fetching its own via database.GetDB(), so it actually runs inside the same serialized transaction #6225 introduced -- previously it sat right next to that race fix without benefiting from it. * fix: address the review findings on the embedded AmneziaWG PR 5 blocking findings: - Floor S3/S4 at 12 in both obfuscation generators (Go and frontend) and reject a hand-edited value below that when HeaderProtectionKey is set -- IpcSet requires it, and ~39% of previously-generated sets violated it silently. - Guard PrivateKey/PrimaryDNS/SecondaryDNS/remark against newline injection in the AmneziaWG .conf builder (both the Go subscription-link path and the frontend downloadable-config path) -- unguarded, any of them could inject an arbitrary config line into a subscriber's client. - Bound the derived AmneziaWG SOCKS relay port to <= 65535 once an inbound's id is known, and check the reverse direction (does the relay port collide with an existing inbound's port) on both create and update -- previously only port -> relay collisions were checked, not relay -> port. - Gate injectAmneziawgV6Egress on the same V6AliasesActive predicate desiredV6Aliases already uses, so the two can't disagree about whether a peer's IPv6 identity is actually active at the OS level. 2 minor findings: - Fix the forwarded-ports cap check's off-by-one (a spec covering exactly the cap was rejected as if it were over it). - Correct docker-compose.yml's stale comment describing the retired DKMS/kernel-module architecture. * chore: retrigger CI build (armv5) failed on a transient Go module proxy network error (INTERNAL_ERROR stream reset on sagernet/sing), unrelated to this PR's changes. * docs: fix doc comments still describing the retired DKMS/awg-quick design A few doc comments (and one illustrative test log line) survived the embedded-architecture cutover unchanged and now contradict the code they sit next to: - internal/amneziawg/types.go's package comment claimed this package still owns a Manager that reconciles OS-level interfaces via awg-quick/DKMS -- that Manager was removed; the reconcile loop lives in internal/amneziawgnet now, and this package is protocol-shape-only. - internal/amneziawg/params.go's ValidateObfuscation/ValidateConfigValue comments cited "awg-quick up" / "awg-quick executes as root" as the reason to validate -- the server itself never calls awg-quick in this architecture; the same value still reaches a real rendered .conf that a client app or an admin's own awg-quick CLI applies downstream, so the validation is still warranted, just for a different consumer. Mirrored the same fix in inbound_amneziawg.go's matching comment and its test's comment. - internal/amneziawgnet/manager.go's Manager doc comments (x3) pointed readers at "internal/amneziawg.Manager" for comparison -- that type no longer exists in this diff at all. Repointed at internal/mtproto.Manager, the pattern this was actually modeled on and the one that's still real. - Swapped one test's illustrative "awg-quick up awg2 failed" log line for a message shaped like this architecture's actual amneziawgnet logging, so a reader skimming the test doesn't wonder whether the server still shells out to awg-quick. No behavior change. * fix(docs): re-run codegen for xray-settings.mdx after conflict merge The automated conflict-resolution hand-merge for this generated file was content-correct but didn't byte-match a real regen (different YAML long-string folding style). Re-ran npm run gen + docs' gen:api and kept that canonical output instead. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix(amneziawg): drop the dead access-log email backfill amneziawgEmailIndex keyed peers by "<tag>|<tunnel IP>", a scheme built for the retired TPROXY bridge where the peer's decapsulated tunnel address survived as the access log's from-address. The embedded architecture relays through a loopback SOCKS5 dial, so every AmneziaWG log line's from-address is 127.0.0.1:<ephemeral> and the lookup could never match: the index was rebuilt on every log view just to miss. Remove the index, its GetXrayLogs wiring and its test. If per-line emails are wanted back, the relay would have to publish a local-port->email registry for the viewer to resolve loopback sources. * fix(api): generate AmneziaWGLogs/PeerActivity schemas instead of hand-writing them The amneziawglogs endpoint's response structs were missing from openapigen's StructAllow, so they were silently absent from every generated schema/example, the endpoints.ts entry carried a hand-written response, and AmneziaWGLogModal.tsx duplicated the shapes as local interfaces - the exact drift the allowlist rule exists to prevent. Allowlist both structs with example tags, point the endpoint at the generated schema, import the generated types in the modal, and sync docs/public/openapi.json. * chore(amneziawg): drop the unreferenced quiccapture package Nothing imports internal/amneziawg/quiccapture and no route exposes it; its package doc justifies the code as a port of frontend/src/lib/xray/i1Generators.ts, which does not exist in this repository, and promises an API round-trip that also does not exist. 1,110 lines of unreachable code with misleading provenance claims. Revert this commit to bring the package back when the live-capture I1 feature and its frontend counterpart actually land. * fix(clients): re-run cross-inbound conflict checks on the serialized writer The new client-level checks - cross-inbound AllowedIPs collisions and AmneziaWG forwardedPorts conflicts - read a fresh DB snapshot, decide, and only then enter runSerializedTx, while lockInbound only serializes writers on the SAME inbound. Two concurrent client creates on two different tunnel inbounds both passed the read and both committed, yielding two peers with one address: the exact check-then-claim race 81cfd857 (#6225) closed for AddInbound, which this PR's own checkAmneziawgnetSocksReverseConflict already cites. Keep the pre-tx pass for fail-fast UX and re-validate inside the transaction, where the single writer makes the answer authoritative. The race test drives two goroutines at two inbounds and demands exactly one winner; it fails with committed=2 when the in-tx re-check is removed. * fix(amneziawg): hot-apply depletion disables like mtproto does applyTrafficMutationBatch special-cases MTProto so a quota/expiry depletion cuts the sidecar immediately, but AmneziaWG fell through to runtime AddUser/RemoveUser - explicit no-ops for this protocol - so a depleted peer kept tunneling until the next 10s reconcile tick. Route it through applyLocalAmneziaWG, whose own contract (re-read committed settings, filter depleted clients, push to the interface) is exactly this case; the comment claiming it mirrors applyLocalMtproto is now true for the depletion path too. * fix(amneziawg): persist cleared DNS fields instead of resurrecting defaults PrimaryDNS/SecondaryDNS marshaled with omitempty, so clearing them persisted settings with no key at all - and the frontend re-parses stored settings through a Zod schema whose .default('8.8.8.8') / .default('8.8.4.4') fire on missing keys, silently repopulating the form on every load and re-persisting the defaults on the next save. Blank is a documented, meaningful state (no DNS line in client configs); drop omitempty so a cleared value survives the round-trip. The regression test normalizes a server block with cleared DNS and fails when the keys are dropped. * fix(amneziawg): accept cleared numeric obfuscation/subnet fields in the form AntD InputNumber emits null when cleared, Zod .default() only replaces undefined, and unlike wireguard.ts - whose optionalClearedInt comment documents exactly this failure mode - the AmneziaWG schema declared subnetCidr and jc/jmin/jmax/s1-s4 as bare z.number() defaults. Clearing any of the eight fields made safeParse reject the null and block the save until the user retyped a value. Absorb null into undefined while keeping each field's schema default, so a cleared field refills its documented default and legacy blobs with absent keys behave as before. * fix(amneziawg): guard the third .conf emitter against newline injection The review-round fix added the newline guard to amneziaWGConfigText (Go) and buildAmneziaWGClientConfig, but genAmneziaWGConfig in inbound-link.ts - the third of the three emitters its own comment says must not drift - still rendered privateKey/primaryDns/secondaryDns/remark unescaped, so a newline there injected a config line (e.g. a rogue PostUp) into the inbound form's downloaded .conf. Add the same guard, plus the regression tests the original fix shipped without: all four fields on the Go and both frontend emitters go red if any guard is removed. * test(amneziawg): pin the S3/S4 floors the TS drift guard claims to mirror The test's docstring says it mirrors internal/amneziawg/params_test.go, but it asserted S3>=8/S4>=4 while the Go test and both generators pin 12/12 - the floor ValidateObfuscation enforces whenever a header protection key is set, which this generator always sets. A regression narrowing the TS floors into 8-11/4-11 would have passed the drift guard and produced configs the backend rejects on save. * docs: restore the pia repo-map entry and document the AmneziaWG subsystem Merging main dropped CLAUDE.md's internal/pia/ bullet (added by #6272) while resolving the repo-map conflict - the package itself is untouched. Restore it, add the missing map entries for the two packages this branch introduces (internal/amneziawg/, internal/amneziawgnet/), bump the cron count, and give amneziawg_job its row in architecture.md's 5.4 table. * chore(amneziawg): correct comments stranded by the architecture pivot ae77c7e9's cutover to the embedded gVisor path deleted the kernel-module code but left several comments describing it in the present tense: hot_diff.go cited the removed service.amneziawgEgressStreamSettings and wrongly claimed AmneziaWG is the only sockopt.tproxy source (tunnel's TProxy mode is the live one the guard protects), socks_config.go pointed at the deleted EgressBasePort/EgressPortForInbound, manager.go referred to the deleted Manager and its fingerprinting as live code, web.go's cron registration claimed the job scrapes traffic (its own doc says it does not), and types.go capped ContentPaddingAddition at uint16 when validation and upstream both use uint32. * style(lint): satisfy gofumpt/goimports so make verify is green json_service.go's two 'Tag: "proxy"}' literals came in with main's own cc245a90 formatting commit and fail the repo's gofumpt gate for everyone; the import grouping in inbound_amneziawg.go is from the serialized-writer fix on this branch. --------- Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com> Co-authored-by: Sanaei <ho3ein.sanaei@gmail.com> Co-authored-by: claude[bot] <41898282+claude[bot]@users.noreply.github.com>
2026-08-23 17:41:15 -07:00
tag_version=$(resolve_latest_tag)
2026-05-19 04:09:35 -07:00
if [[ ! -n "$tag_version" ]]; then
echo -e "${red}Failed to fetch x-ui version, it may be due to GitHub API restrictions, please try it later${plain}"
exit 1
2026-05-19 04:09:35 -07:00
fi
echo -e "Got x-ui latest version: ${tag_version}, beginning the installation..."
curl -fLR --retry 5 --retry-delay 3 --connect-timeout 15 --speed-limit 1 --speed-time 300 -o ${xui_folder}-linux-$(arch).tar.gz https://github.com/MHSanaei/3x-ui/releases/download/${tag_version}/x-ui-linux-$(arch).tar.gz
2026-05-19 04:09:35 -07:00
if [[ $? -ne 0 ]]; then
echo -e "${red}Downloading x-ui failed, please be sure that your server can access GitHub ${plain}"
exit 1
fi
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if [[ ! -s ${xui_folder}-linux-$(arch).tar.gz ]]; then
rm ${xui_folder}-linux-$(arch).tar.gz -f
echo -e "${red}Downloaded x-ui release archive is empty${plain}"
exit 1
fi
verify_release_checksum "https://github.com/MHSanaei/3x-ui/releases/download/${tag_version}/x-ui-linux-$(arch).tar.gz" "${xui_folder}-linux-$(arch).tar.gz"
2026-05-19 04:09:35 -07:00
else
tag_version=$1
# The rolling dev channel ships under a fixed, non-semver tag that is
# force-moved to the latest main commit on every push. Accept `dev` as a
# convenient alias and skip the numeric floor check for it.
if [[ "$tag_version" == "dev" || "$tag_version" == "dev-latest" ]]; then
tag_version="dev-latest"
echo -e "${yellow}Installing the rolling dev build (tag: dev-latest). This is a per-commit pre-release, not a stable version.${plain}"
else
tag_version_numeric=${tag_version#v}
min_version="2.3.5"
2026-05-19 04:09:35 -07:00
if [[ "$(printf '%s\n' "$min_version" "$tag_version_numeric" | sort -V | head -n1)" != "$min_version" ]]; then
echo -e "${red}Please use a newer version (at least v2.3.5). Exiting installation.${plain}"
exit 1
fi
2026-05-19 04:09:35 -07:00
fi
url="https://github.com/MHSanaei/3x-ui/releases/download/${tag_version}/x-ui-linux-$(arch).tar.gz"
echo -e "Beginning to install x-ui ${tag_version}"
curl -fLR --retry 5 --retry-delay 3 --connect-timeout 15 --speed-limit 1 --speed-time 300 -o ${xui_folder}-linux-$(arch).tar.gz ${url}
2026-05-19 04:09:35 -07:00
if [[ $? -ne 0 ]]; then
echo -e "${red}Download x-ui ${tag_version} failed, please check if the version exists ${plain}"
2026-05-19 04:09:35 -07:00
exit 1
fi
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if [[ ! -s ${xui_folder}-linux-$(arch).tar.gz ]]; then
rm ${xui_folder}-linux-$(arch).tar.gz -f
echo -e "${red}Downloaded x-ui release archive is empty${plain}"
exit 1
fi
verify_release_checksum "${url}" "${xui_folder}-linux-$(arch).tar.gz"
2026-05-19 04:09:35 -07:00
fi
fix(install): fetch x-ui.sh and unit files from the installed release tag (#6391) * fix(install): fetch x-ui.sh and unit files from the installed release tag install.sh and update.sh pin the panel archive to a release tag but always took x-ui.sh, x-ui.rc and the service units from main, so the management script and the binary of one installation came from different commits: the fail2ban templates and setting flags the script writes drift silently against an older binary, two installs of the same tag differ, and a reviewed or digest-pinned installer still runs unreviewed code from main. Use the same ref as the archive, keeping main only for the rolling dev-latest build. The menu's "update menu" and update_shell paths now fetch the script matching the installed version and fall back to main with a visible notice when no script is published for it. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): fall back to main for files a pinned tag does not publish Review follow-ups. install.sh accepts tags down to v2.3.5, but x-ui.rc only exists from v2.8.4 and the split x-ui.service.* files are newer still, so pinning those to the tag made an Alpine install of an old tag 404 after the previous install was already removed. Probe the tag for each file and fall back to main with a notice when it is missing, as the menu already does for x-ui.sh. The fail2ban auto-setup probe also trusted the exit status of 'x-ui setup-fail2ban', but scripts before v3.4.0 have no such subcommand and exit 0 from the usage banner, so the installer reported a setup that never ran. Skip with a notice when the installed script does not know the subcommand. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): refuse a tag that does not publish a needed script Falling back to main reintroduced the binary/script mismatch the tag pinning exists to remove, and it fired at points where install.sh and update.sh have already stopped and removed the previous installation -- so the quiet path was also the one that could not be undone. Probe the tag instead, before anything is touched, for every file that is always fetched from GitHub (x-ui.sh, plus x-ui.rc on Alpine), and abort with the HTTP status when one is missing. The unit files stay unprobed: they are only fetched when the release tarball omits them, so an old tag that ships x-ui.service inside its tarball still installs. Their existing failure message now names the ref it tried. Also tighten the setup-fail2ban probe to the dispatcher's case arm rather than any mention of the string, which also matches a comment. --------- Co-authored-by: Claude <noreply@anthropic.com>
2026-09-03 07:50:53 -07:00
# x-ui.sh, x-ui.rc and the unit files must come from the same release as
# the binary; only the rolling dev build tracks main.
local script_ref="${tag_version}"
if [[ "${tag_version}" == "dev-latest" ]]; then
script_ref="main"
fi
# The unit files are only fetched when the release tarball lacks them, so
# they are checked at that point instead of here.
local required_files=("x-ui.sh")
[[ $release == "alpine" ]] && required_files+=("x-ui.rc")
require_repo_files "${script_ref}" "${required_files[@]}"
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
local xui_script_temp="/usr/bin/x-ui-temp.$$"
rm -f "${xui_script_temp}"
fix(install): fetch x-ui.sh and unit files from the installed release tag (#6391) * fix(install): fetch x-ui.sh and unit files from the installed release tag install.sh and update.sh pin the panel archive to a release tag but always took x-ui.sh, x-ui.rc and the service units from main, so the management script and the binary of one installation came from different commits: the fail2ban templates and setting flags the script writes drift silently against an older binary, two installs of the same tag differ, and a reviewed or digest-pinned installer still runs unreviewed code from main. Use the same ref as the archive, keeping main only for the rolling dev-latest build. The menu's "update menu" and update_shell paths now fetch the script matching the installed version and fall back to main with a visible notice when no script is published for it. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): fall back to main for files a pinned tag does not publish Review follow-ups. install.sh accepts tags down to v2.3.5, but x-ui.rc only exists from v2.8.4 and the split x-ui.service.* files are newer still, so pinning those to the tag made an Alpine install of an old tag 404 after the previous install was already removed. Probe the tag for each file and fall back to main with a notice when it is missing, as the menu already does for x-ui.sh. The fail2ban auto-setup probe also trusted the exit status of 'x-ui setup-fail2ban', but scripts before v3.4.0 have no such subcommand and exit 0 from the usage banner, so the installer reported a setup that never ran. Skip with a notice when the installed script does not know the subcommand. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): refuse a tag that does not publish a needed script Falling back to main reintroduced the binary/script mismatch the tag pinning exists to remove, and it fired at points where install.sh and update.sh have already stopped and removed the previous installation -- so the quiet path was also the one that could not be undone. Probe the tag instead, before anything is touched, for every file that is always fetched from GitHub (x-ui.sh, plus x-ui.rc on Alpine), and abort with the HTTP status when one is missing. The unit files stay unprobed: they are only fetched when the release tarball omits them, so an old tag that ships x-ui.service inside its tarball still installs. Their existing failure message now names the ref it tried. Also tighten the setup-fail2ban probe to the dispatcher's case arm rather than any mention of the string, which also matches a comment. --------- Co-authored-by: Claude <noreply@anthropic.com>
2026-09-03 07:50:53 -07:00
curl -fLRo "${xui_script_temp}" "https://raw.githubusercontent.com/MHSanaei/3x-ui/${script_ref}/x-ui.sh"
2026-05-19 04:09:35 -07:00
if [[ $? -ne 0 ]]; then
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
rm -f "${xui_script_temp}"
2026-05-19 04:09:35 -07:00
echo -e "${red}Failed to download x-ui.sh${plain}"
exit 1
fi
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if [[ ! -s "${xui_script_temp}" ]]; then
rm -f "${xui_script_temp}"
echo -e "${red}Downloaded x-ui.sh is empty${plain}"
exit 1
fi
2026-05-19 04:09:35 -07:00
# Stop x-ui service and remove old resources
fix(install): preserve custom bin/ files (e.g. hand-added geoip) across updates (#6152) * fix(install): preserve custom bin/ files (e.g. hand-added geoip) across updates Every reinstall/update wipes /usr/local/x-ui/ wholesale and re-extracts the release tarball, which only ships known assets (xray/mtg binaries, the bundled geoip*/geosite*.dat sets). A user-reported real incident: a hand-placed custom geoip file referenced from a routing rule via "ext:<file>:<code>" got silently deleted on update, and Xray refused to start at all afterward ("failed to open <file>: no such file or directory"), taking down every inbound until the file was manually restored from the user's own backup. install_x-ui now backs up the old bin/ before the wipe and restores, after extraction, only the files the fresh release doesn't provide -- bundled assets still get the newer per-release copy, nothing custom silently disappears. Verified in isolation: standard files (geoip.dat, the xray binary) end up as the fresh release's copy; a custom file absent from the release survives untouched. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix: harden the bin/ snapshot-and-restore against the review round on #6152 - Replace the mktemp+cp snapshot with a same-filesystem mv of bin/ aside: an unchecked mktemp failure previously made the very next line copy bin/'s contents into "/" (empty custom_bin_backup + trailing slash), and a silently-ignored cp failure (stderr redirected, exit code never checked) could leave a truncated custom geo file that gets "restored" as if it were intact. A rename is atomic and needs no extra disk space, removing both failure modes at once; if it fails, back off cleanly and say so instead of proceeding as if a backup exists. - Add a trap so an interrupted update (Ctrl-C, signal) between the backup and the restore doesn't leave the snapshot (which contains bin/config.json and every mtproto client's FakeTLS secret) sitting around indefinitely; the two exit-path cleanups this replaces are gone since the trap now covers those exits too. - Move the restore below the arm arch-rename/chmod block instead of before it, so xray-linux-arm32/mtg-linux-arm already exist under their final names and don't get needlessly restored-then-overwritten and misreported as "custom". - Exclude bin/config.json and bin/mtproto/*.toml from the restore: those are the panel's own generated runtime state (internal/xray/process.go, internal/mtproto/manager.go), not admin-placed files, and restoring a stale one only resurrects dead state or recreates bin/mtproto/ with the wrong (more permissive) directory mode. - Match symlinks in the restore's find, not just plain files -- cp -a already preserves them in the snapshot, but the restore loop was silently dropping them, which is exactly the failure mode (a geo file symlinked in from elsewhere) this PR set out to fix. - Quote the two new xui_folder expansions. - Extend the non-interactive smoke test to reinstall over an existing install with a sentinel file in bin/, asserting it survives and that the bundled geoip.dat is still the release's own copy -- the update path this PR touches had no CI coverage at all before this. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-14 07:41:45 -07:00
local custom_bin_backup=""
2026-05-19 04:09:35 -07:00
if [[ -e ${xui_folder}/ ]]; then
if [[ $release == "alpine" ]]; then
rc-service x-ui stop
else
systemctl stop x-ui
fi
fix(mtproto): reap orphaned mtg, fix SysLog viewer, mtg log visibility, export remark (#5105) (#5107) * fix(logs): render journalctl output in the SysLog viewer The log viewer's parseLogLine only understood the app-log format (2006/01/02 15:04:05 LEVEL - body). With SysLog ticked the backend returns journalctl lines (Mon DD HH:MM:SS host ident[pid]: LEVEL - body), so the parser mistook the journal time for the level and dropped the body, leaving only timestamps. Detect and strip the journald prefix, keep the journal timestamp as the stamp, then parse the real level and body from the remainder. * feat(mtproto): surface mtg output and add status reporting mtg's stdout/stderr was captured by a writer that kept only the last line and showed it nowhere, so the reason a proxy could not reach Telegram was invisible. Stream mtg output line-by-line into the x-ui log, tagged per inbound, so it appears in the panel log viewer and journald. Also fix mangled log lines: logger.Info uses fmt.Sprint, which drops the space between adjacent string operands, producing output like 'inbound3on0.0.0.0:8443'. Switch the affected mtproto calls to the formatted (*f) variants. Add show_mtproto_status to x-ui.sh so 'x-ui status' reports each mtproto inbound's mtg process state and bind address. * fix(logs): parse all journalctl message shapes in SysLog viewer Real journalctl output mixes four message shapes after the 'Mon DD HH:MM:SS host ident[pid]:' prefix: go-logging 'LEVEL - msg' (x-ui/xray), Go std-log with an embedded date (net/http, runtime), telego's '[timestamp] LEVEL msg', and systemd lines. The viewer only understood the first, so std-log and telego lines — which never contain ' - ' — collapsed to a bare timestamp (e.g. the 8s telego 409 spam). Extract the parser into a pure, testable module and teach it the other shapes: strip the redundant Go std-log date, lift the level out of telego brackets, and always keep the message body. Add a unit test covering each shape with real captured lines. * fix(mtproto): reap orphaned mtg sidecars so a stale one can't break new clients On Linux x-ui does not kill its mtg children when it dies (no kill-on-exit, unlike the Windows job object). After a crash, OOM, kill -9, or update, a stale mtg keeps holding the inbound port with an OLD secret, so new clients fail the FakeTLS handshake and get silently domain-fronted to the fakeTLS domain instead of proxied to Telegram (a few MB of traffic, never connects). Sweep orphans at startup: on the first reconcile, before x-ui starts any of its own mtg, scan /proc and SIGKILL any process whose executable is our mtg-<goos>-<goarch> binary. x-ui is the sole owner of mtg, so anything alive then is an orphan. Runs once per process (swept guard), survives the binary-deleted-during-update case via /proc/<pid>/cmdline, and is a no-op on Windows (job object) and other platforms. Also clear stray mtg in update.sh/install.sh after stopping x-ui, anchored to the 'mtg-linux-<arch> run ' invocation so the pattern can't match unrelated command lines (e.g. x-ui.sh's own 'grep mtg-linux'). * fix(logs): drop dead body initializer flagged by eslint no-useless-assignment * fix(mtproto): drop remark fragment from tg://proxy export link The mtproto export link appended the inbound remark as a URL fragment (tg://proxy?server=...&port=...&secret=...#remark). Telegram Desktop rejects a proxy deep link with a trailing fragment as 'This proxy link is invalid', breaking one-click import, and a remark is meaningless for proxy links across clients. Stop adding it in both the panel link (genMtprotoLink) and the subscription service. Fixes #5105. * fix(x-ui.sh): remove unused check_mtproto_status helper show_mtproto_status does its own process check, so check_mtproto_status was dead code. Drop it (per Copilot review on #5107).
2026-06-08 19:01:33 -07:00
# Kill any leftover mtg (MTProto) sidecars. x-ui runs them outside its own
# lifecycle, so on Linux a stale one can survive the stop and keep holding
# an inbound port with an outdated secret, silently breaking new clients.
# The freshly installed panel respawns a clean mtg per inbound on start.
pkill -f 'mtg-linux-[^ ]* run ' > /dev/null 2>&1 || true
Feature/tuic v5 (#6337) * Feat(tuic): Implement native TUIC v5 protocol support via Rust sidecar daemon - Add internal/tuic package for official tuic-server sidecar lifecycle management, configuration generation, and graceful process control - Bridge decrypted TUIC QUIC traffic into loopback Xray SOCKS5 inbounds (63200+id) for traffic accounting, statistics, and routing rules - Implement periodic reconciliation job (cadence @every 10s) and immediate runtime synchronization on inbound/client mutations - Add TUIC inbound & multi-user client settings (UUID + Password authentication) in Web UI with SNI auto-fill and panel certificate loader - Integrate tuic:// subscription links and Clash.Meta (Mihomo) proxy generation for TUIC - Update install.sh to automatically download and install official tuic-server release for x86_64, aarch64, and armv7 - Add full localization for TUIC protocol across all 13 supported languages * Feat(install): Support custom repository and branch in install and update scripts * Ci(release): Enable publish-dev for feature branch and workflow dispatch * Feat(sub): Add TUIC to subscription resolution and client QR config generator - Add 'tuic' to getInboundsBySubId SQL allowlist to resolve TUIC inbounds in subscriptions and sub links - Enhance buildTuicProxy in Clash subscription generator with robust host and credentials resolution - Add tuicConfig.ts to generate standalone Clash/Mihomo YAML configuration - Add dedicated TUIC Config tab in ClientQrModal with QR code and .yaml download button - Add localization keys for TUIC config across all 13 supported languages * Fix(tuic): Exclude TUIC from native Xray inbounds and strip udp_relay_mode from server config - Exclude model.TUIC from native Xray inbounds in GetXrayConfig to prevent Xray startup failure - Remove udp_relay_mode from tuic-server JSON configuration builder - Update install.sh to install tuic-server binary to both xui_folder/bin and /usr/local/bin * Fix(install): Fallback to dev-latest when releases/latest is not present on fork * Feat(tuic): Add real-time online status and LastOnline tracking for TUIC clients - Track client activity by mapping client UUID in tuic-server logs to email - Integrate TUIC active clients into XrayTrafficJob to refresh local online clients - Bump LastOnline timestamp in database and broadcast live online status over WebSocket * Feat(tuic): Implement real-time traffic statistics and live speed reporting for TUIC - Collect precise I/O traffic deltas for tuic-server child processes via /proc/<pid>/io - Aggregate and attribute TUIC traffic deltas per client in tuic Manager - Integrate TUIC traffic deltas into XrayTrafficJob to update database and broadcast live speed * Feat(tuic): Finalize TUIC v5 integration with 1:1 traffic counting and orphan process cleanup - Use exact 1:1 byte delta accounting from /proc/<pid>/io - Add killStrayTuicProcesses to terminate orphan sidecars on panel startup - Fully integrate TUIC with subscriptions, live speed meter, and all 13 locales * Feat(frontend): Polish TUIC UI, support bulk operations, and update translations - Align TUIC inbound certificate form with standard 3X-UI layout (Set Default Cert, Clear) - Remove extra subtitle hint text from TUIC inbound form fields - Support TUIC in client bulk attach/detach and bulk add modals - Add TUIC badge color to client info modal, clients table, and host list - Update password tooltip across all 13 locales to include TUIC - Remove obsolete dead translation keys across all 13 locales * Chore(ci): Finalize TUIC v5 bundling across release workflow, Docker, and scripts * Feat(openapi): Update OpenAPI generator and schemas for TUIC types * Fix(backend): Address core review findings for TUIC types, port checks, and xray bridge * Refactor(traffic): Isolate proc reading with build tags and decouple TUIC metering into TuicJob * Feat(client): Add TuicServer to InboundOption, fix config export and clean share links * Fix(frontend): Register TUIC in multi-user helpers, tracked protocols, and tag derivation * Chore(openapi): Re-generate OpenAPI specification and sync Zod schemas * Chore(scripts): Add Alpine musl binaries, 386 and Windows packaging, and anchor pkill * Fix(review): Remove stale import, correct binary names, switch to musl, and drop unreachable relay gate * Feat(frontend): Show share link in Inbound Info and display UDP tag for TUIC * Docs: Add TUIC v5 configuration guide and link specifications * Docs(tuic): Correct Clash Meta configuration parameter to reduce-rtt * Fix(tuic): Generate client credentials on copy, enforce ID/password validation, and add i386 to DockerInit * Fix(tuic): drop unused relay, fix traffic accounting, and honor host endpoints - Drop unused loopback SOCKS relay and eliminate port collision with AmneziaWG - Correct inbound traffic calculation without double-counting - Drop heuristic client traffic division while retaining online tracking - Support externalProxy host fan-out and conditional parameters in share links - Scope orphan process termination to managed config directory * Fix(tuic): enforce client quotas, decouple Xray restart, and sync openapi schemas - Regenerate OpenAPI, Zod schemas, and TypeScript types without route_through_xray - Populate clientTraffics in TuicJob to enforce client quotas and first-use expiry - Split process I/O delta into up and down in Process.CollectTraffic - Remove SetNeedRestart from updateTuicInbound to prevent Xray session drops - Use InstanceFromInbound for default ALPN and UDP relay mode in tuic:// share links - Support allow_insecure on externalProxy host endpoints without parameter collision * Fix(tuic): attribute client traffic only on single-user inbounds and sync link defaults - Attribute I/O deltas to the client only when the inbound has exactly one configured client, avoiding false billing and disablings on multi-user inbounds - Aggregate client traffic by email in TuicJob so clients on multiple inbounds don't lose deltas - Match frontend genTuicLink defaults for alpn and udp_relay_mode with backend subscription links * Fix(tuic): gate client traffic by total sidecar clients and require client email * Fix(tuic): enforce inbound-only traffic limits and disable client totalGB * fix(tuic): restore delayed start, remove client totalGB rejection, and document linux-only limits * fix(tuic): anchor pkill, fix io baseline/split, escape yaml, and deduplicate start errors * fix(tuic): prevent traffic double-counting, ensure info log level for delayed start, and broaden pkill matching * fix(tuic): address review round 11 findings - internal/sub/json_service: skip tuic protocol in json subscription to prevent direct routing leak - internal/sub/clash_service: honor externalProxy/host row allowInsecure, sni, and alpn in buildTuicProxy - internal/web/runtime: decouple tuic inbound add/delete from xray restart - internal/tuic/config: restore user log-level options (warn, error) without forced info clamp - frontend/src/lib/xray/inbound-link: fix duplicate remark suffix and apply externalProxy TLS overrides - frontend/src/schemas/protocols/stream/external-proxy: propagate allowInsecure through host mapping - tests: add coverage for json sub skip, clash proxy overrides, and link generation * fix(tuic): meter inbound traffic through a UDP relay and bracket IPv6 binds Review repairs on the TUIC v5 sidecar integration: - Inbound traffic was read from the sidecar's /proc/<pid>/io rchar, but the kernel only counts read()/write() there and tuic-server moves its sockets with recvfrom/recvmmsg/sendmmsg/sendto, so an inbound's up/down stayed at 0 forever and inbound total limits never tripped (measured: 12 MiB relayed, rchar delta 0). The panel now owns the inbound's public UDP port with a small relay and runs tuic-server behind it on a loopback port, counting up/down exactly on every OS. tuic-server therefore logs 127.0.0.1 as every client's address; per-client attribution stays unsupported since QUIC is opaque. - Instance.BindTo formatted an IPv6 listen address as ":::8443", which tuic-server rejects with "invalid socket address syntax", so an inbound listening on "::" or any IPv6 literal never started. It now uses net.JoinHostPort; IPv4 output is unchanged. - The log level is passed to the sidecar as chosen. Online status, last-online and delayed start are read from its Info lines, so the Log Level field now says that Warn and Error switch them off for the inbound, and the docs say the same. - Drop two frontend tests that only exercised a getter and a set lookup, and strip the trailing blank line that made gofumpt fail on two of the new Go test files. * fix(tuic): harden tag updates, runtime routing, and relay stability --------- Co-authored-by: poise52 <equipoise52@gmail.com> Co-authored-by: Sanaei <ho3ein.sanaei@gmail.com>
2026-09-12 01:15:48 -07:00
pkill -f 'tuic-server.*-c .*bin/tuic/tuic_[0-9]+\.json' > /dev/null 2>&1 || true
fix(install): preserve custom bin/ files (e.g. hand-added geoip) across updates (#6152) * fix(install): preserve custom bin/ files (e.g. hand-added geoip) across updates Every reinstall/update wipes /usr/local/x-ui/ wholesale and re-extracts the release tarball, which only ships known assets (xray/mtg binaries, the bundled geoip*/geosite*.dat sets). A user-reported real incident: a hand-placed custom geoip file referenced from a routing rule via "ext:<file>:<code>" got silently deleted on update, and Xray refused to start at all afterward ("failed to open <file>: no such file or directory"), taking down every inbound until the file was manually restored from the user's own backup. install_x-ui now backs up the old bin/ before the wipe and restores, after extraction, only the files the fresh release doesn't provide -- bundled assets still get the newer per-release copy, nothing custom silently disappears. Verified in isolation: standard files (geoip.dat, the xray binary) end up as the fresh release's copy; a custom file absent from the release survives untouched. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix: harden the bin/ snapshot-and-restore against the review round on #6152 - Replace the mktemp+cp snapshot with a same-filesystem mv of bin/ aside: an unchecked mktemp failure previously made the very next line copy bin/'s contents into "/" (empty custom_bin_backup + trailing slash), and a silently-ignored cp failure (stderr redirected, exit code never checked) could leave a truncated custom geo file that gets "restored" as if it were intact. A rename is atomic and needs no extra disk space, removing both failure modes at once; if it fails, back off cleanly and say so instead of proceeding as if a backup exists. - Add a trap so an interrupted update (Ctrl-C, signal) between the backup and the restore doesn't leave the snapshot (which contains bin/config.json and every mtproto client's FakeTLS secret) sitting around indefinitely; the two exit-path cleanups this replaces are gone since the trap now covers those exits too. - Move the restore below the arm arch-rename/chmod block instead of before it, so xray-linux-arm32/mtg-linux-arm already exist under their final names and don't get needlessly restored-then-overwritten and misreported as "custom". - Exclude bin/config.json and bin/mtproto/*.toml from the restore: those are the panel's own generated runtime state (internal/xray/process.go, internal/mtproto/manager.go), not admin-placed files, and restoring a stale one only resurrects dead state or recreates bin/mtproto/ with the wrong (more permissive) directory mode. - Match symlinks in the restore's find, not just plain files -- cp -a already preserves them in the snapshot, but the restore loop was silently dropping them, which is exactly the failure mode (a geo file symlinked in from elsewhere) this PR set out to fix. - Quote the two new xui_folder expansions. - Extend the non-interactive smoke test to reinstall over an existing install with a sentinel file in bin/, asserting it survives and that the bundled geoip.dat is still the release's own copy -- the update path this PR touches had no CI coverage at all before this. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-14 07:41:45 -07:00
# bin/ is about to be wiped wholesale by the tar extraction below. The
# release only ships known assets (xray/mtg binaries, the bundled
# geoip*/geosite*.dat sets) -- anything else in bin/ was placed there
# by the admin (e.g. a hand-added custom geoip/geosite file referenced
# from a routing rule via "ext:<file>:<code>") and would otherwise be
# silently deleted on every update, breaking Xray at next start with
# "failed to open <file>: no such file or directory" for any routing
# rule that references it. Moved aside rather than copied: a rename
# on the same filesystem is atomic (no truncated file if disk space
# runs out mid-copy, unlike `cp`) and keeps the snapshot under
# /usr/local rather than a separate, possibly small/tmpfs $TMPDIR.
if [[ -d "${xui_folder}/bin" ]]; then
custom_bin_backup="${xui_folder%/x-ui}/x-ui-bin-backup.$$"
rm -rf "${custom_bin_backup}"
if ! mv "${xui_folder}/bin" "${custom_bin_backup}"; then
custom_bin_backup=""
echo -e "${yellow}Could not back up bin/ -- custom files there will not be preserved across this update${plain}"
fi
fi
# Sole cleanup path for the backup from here on -- covers both the
# two `exit 1`s below (extraction/binary-missing failures) and an
# interrupted update (Ctrl-C, signal) before the restore runs.
# Cleared once the restore below finishes normally.
trap '[[ -n "${custom_bin_backup}" ]] && rm -rf "${custom_bin_backup}"' EXIT INT TERM
2026-05-19 04:09:35 -07:00
rm ${xui_folder}/ -rf
fi
# Extract resources and set permissions
tar zxvf x-ui-linux-$(arch).tar.gz
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if [[ $? -ne 0 ]]; then
rm x-ui-linux-$(arch).tar.gz -f
rm -f "${xui_script_temp}"
echo -e "${red}Failed to extract the x-ui release archive -- the previous installation has already been removed, so the panel will not start until this is fixed; try running the installer again${plain}"
exit 1
fi
2026-05-19 04:09:35 -07:00
rm x-ui-linux-$(arch).tar.gz -f
cd x-ui
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if [[ $? -ne 0 || ! -s x-ui ]]; then
rm -f "${xui_script_temp}"
echo -e "${red}Extracted x-ui archive is missing the x-ui binary -- the previous installation has already been removed, so the panel will not start until this is fixed; try running the installer again${plain}"
exit 1
fi
2026-05-19 04:09:35 -07:00
chmod +x x-ui
chmod +x x-ui.sh
2026-05-04 04:20:24 -07:00
# Check the system's architecture and rename the file accordingly.
# The panel binary maps GOARCH=arm to "arm32" (internal/xray/process.go),
# so the Xray binary must be named xray-linux-arm32; mtg keeps plain "arm".
2024-04-01 01:42:28 -07:00
if [[ $(arch) == "armv5" || $(arch) == "armv6" || $(arch) == "armv7" ]]; then
mv bin/xray-linux-$(arch) bin/xray-linux-arm32
chmod +x bin/xray-linux-arm32
feat(mtproto): add MTProto (FakeTLS) protocol via managed mtg sidecar (#5076) * feat(mtproto): add MTProto (FakeTLS) protocol via managed mtg sidecar Xray-core has no mtproto proxy, so mtproto inbounds run as standalone mtg (9seconds/mtg) sidecar processes managed by the panel — one per inbound — and are excluded from the generated Xray config entirely. - model: MTProto protocol constant, validator, and FakeTLS secret helpers (GenerateFakeTLSSecret/HealMtprotoSecret) - mtproto package: per-inbound mtg process manager with reconcile, graceful stop, and best-effort Prometheus traffic scraping - runtime: delegate mtproto inbounds to the mtg manager instead of the Xray gRPC API; skip mtproto when building the Xray config - web: boot reconcile + StopAll wiring, periodic reconcile/traffic job, port-conflict transport, secret healing on inbound add/update - sub: tg:// proxy share-link generation - frontend: protocol option, Zod schema, Protocol tab (FakeTLS domain + regenerable secret), info-modal link, and i18n - provisioning: fetch mtg v2.2.8 in install.sh, DockerInit.sh, and the Linux + Windows release workflows * fix * fix * fix: address Copilot review comments on mtproto PR - web/web.go: create NewMtprotoJob once and reuse for cron + initial run - mtproto/manager.go: StopAll cleans up per-inbound config files on shutdown - mtproto/manager.go: CollectTraffic releases mutex before HTTP scrapes to avoid blocking Ensure/Reconcile/Remove during network I/O - database/model/model.go: panic on crypto/rand failure in mtprotoRandomMiddle instead of silently producing a weak all-zero secret - install.sh: fix chmod to handle renamed bin/mtg-linux-arm on armv5/v6/v7
2026-06-08 05:28:19 -07:00
if [[ -f bin/mtg-linux-$(arch) ]]; then
mv bin/mtg-linux-$(arch) bin/mtg-linux-arm
chmod +x bin/mtg-linux-arm
fi
2024-01-15 16:01:15 -08:00
fi
2026-05-19 04:09:35 -07:00
chmod +x x-ui bin/xray-linux-$(arch)
feat(mtproto): add MTProto (FakeTLS) protocol via managed mtg sidecar (#5076) * feat(mtproto): add MTProto (FakeTLS) protocol via managed mtg sidecar Xray-core has no mtproto proxy, so mtproto inbounds run as standalone mtg (9seconds/mtg) sidecar processes managed by the panel — one per inbound — and are excluded from the generated Xray config entirely. - model: MTProto protocol constant, validator, and FakeTLS secret helpers (GenerateFakeTLSSecret/HealMtprotoSecret) - mtproto package: per-inbound mtg process manager with reconcile, graceful stop, and best-effort Prometheus traffic scraping - runtime: delegate mtproto inbounds to the mtg manager instead of the Xray gRPC API; skip mtproto when building the Xray config - web: boot reconcile + StopAll wiring, periodic reconcile/traffic job, port-conflict transport, secret healing on inbound add/update - sub: tg:// proxy share-link generation - frontend: protocol option, Zod schema, Protocol tab (FakeTLS domain + regenerable secret), info-modal link, and i18n - provisioning: fetch mtg v2.2.8 in install.sh, DockerInit.sh, and the Linux + Windows release workflows * fix * fix * fix: address Copilot review comments on mtproto PR - web/web.go: create NewMtprotoJob once and reuse for cron + initial run - mtproto/manager.go: StopAll cleans up per-inbound config files on shutdown - mtproto/manager.go: CollectTraffic releases mutex before HTTP scrapes to avoid blocking Ensure/Reconcile/Remove during network I/O - database/model/model.go: panic on crypto/rand failure in mtprotoRandomMiddle instead of silently producing a weak all-zero secret - install.sh: fix chmod to handle renamed bin/mtg-linux-arm on armv5/v6/v7
2026-06-08 05:28:19 -07:00
if [[ -f bin/mtg-linux-arm ]]; then
chmod +x bin/mtg-linux-arm
elif [[ -f bin/mtg-linux-$(arch) ]]; then
chmod +x bin/mtg-linux-$(arch)
fi
2026-05-04 04:20:24 -07:00
fix(install): preserve custom bin/ files (e.g. hand-added geoip) across updates (#6152) * fix(install): preserve custom bin/ files (e.g. hand-added geoip) across updates Every reinstall/update wipes /usr/local/x-ui/ wholesale and re-extracts the release tarball, which only ships known assets (xray/mtg binaries, the bundled geoip*/geosite*.dat sets). A user-reported real incident: a hand-placed custom geoip file referenced from a routing rule via "ext:<file>:<code>" got silently deleted on update, and Xray refused to start at all afterward ("failed to open <file>: no such file or directory"), taking down every inbound until the file was manually restored from the user's own backup. install_x-ui now backs up the old bin/ before the wipe and restores, after extraction, only the files the fresh release doesn't provide -- bundled assets still get the newer per-release copy, nothing custom silently disappears. Verified in isolation: standard files (geoip.dat, the xray binary) end up as the fresh release's copy; a custom file absent from the release survives untouched. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix: harden the bin/ snapshot-and-restore against the review round on #6152 - Replace the mktemp+cp snapshot with a same-filesystem mv of bin/ aside: an unchecked mktemp failure previously made the very next line copy bin/'s contents into "/" (empty custom_bin_backup + trailing slash), and a silently-ignored cp failure (stderr redirected, exit code never checked) could leave a truncated custom geo file that gets "restored" as if it were intact. A rename is atomic and needs no extra disk space, removing both failure modes at once; if it fails, back off cleanly and say so instead of proceeding as if a backup exists. - Add a trap so an interrupted update (Ctrl-C, signal) between the backup and the restore doesn't leave the snapshot (which contains bin/config.json and every mtproto client's FakeTLS secret) sitting around indefinitely; the two exit-path cleanups this replaces are gone since the trap now covers those exits too. - Move the restore below the arm arch-rename/chmod block instead of before it, so xray-linux-arm32/mtg-linux-arm already exist under their final names and don't get needlessly restored-then-overwritten and misreported as "custom". - Exclude bin/config.json and bin/mtproto/*.toml from the restore: those are the panel's own generated runtime state (internal/xray/process.go, internal/mtproto/manager.go), not admin-placed files, and restoring a stale one only resurrects dead state or recreates bin/mtproto/ with the wrong (more permissive) directory mode. - Match symlinks in the restore's find, not just plain files -- cp -a already preserves them in the snapshot, but the restore loop was silently dropping them, which is exactly the failure mode (a geo file symlinked in from elsewhere) this PR set out to fix. - Quote the two new xui_folder expansions. - Extend the non-interactive smoke test to reinstall over an existing install with a sentinel file in bin/, asserting it survives and that the bundled geoip.dat is still the release's own copy -- the update path this PR touches had no CI coverage at all before this. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-14 07:41:45 -07:00
# Restore anything from the old bin/ that the fresh release doesn't ship
# (custom geoip/geosite files, or anything else an admin hand-placed
# there) -- never overwrites a same-named file the new release provides,
# so bundled assets (geoip.dat, geoip_RU.dat, ...) still get the fresh
# per-release copy. Runs after the arch-rename above so xray-linux-arm32/
# mtg-linux-arm already exist under their final names there and aren't
# mistaken for custom files needing a restore. Skips paths the panel
# itself regenerates at runtime (config.json, mtproto/*.toml -- see
# internal/xray/process.go, internal/mtproto/manager.go): those aren't
# admin-placed, and restoring a stale one only resurrects dead state (an
# orphaned mtg config for a since-deleted inbound) or the wrong
# directory permissions.
if [[ -n "${custom_bin_backup}" ]]; then
local restored_custom_bin=()
while IFS= read -r -d '' f; do
local rel="${f#"${custom_bin_backup}"/}"
case "${rel}" in
feat(tuic): implement native in-process Go TUIC v5 server (#6577) * feat(tuic): implement native in-process Go TUIC v5 server - Implement native TUIC v5 protocol server on pure Go using quic-go - Bridge decrypted TCP/UDP traffic into Xray-core via loopback SOCKS5 inbound - Support full Xray routing rules (geosite/geoip) and cascading outbounds - Implement atomic per-client traffic accounting with TotalGB and ExpiryTime - Add automatic legacy cleanup for older Rust tuic-server binaries, configs, and orphaned processes - Eliminate external Rust tuic-server downloads from install/CI scripts * fix(tuic): address traffic accounting, client reload, and socket lifecycle issues * fix(service): update checkTuicSocksReverseConflict to use bindAddr for listenOverlaps * fix(tuic): resolve traffic double-accounting, UDP fragmentation, and socket lifecycle issues * feat(tuic): complete native Go integration and address audit findings - Integrate an isolated QUIC fork pinned to a specific commit - Preserve original QUIC dependencies for Xray, Hysteria and Gin - Apply BBR, CUBIC and Reno to server connections and exported client profiles - Bridge Xray BBR with correct monotonic time and congestion type conversions - Handle congestion sender recreation after PMTU changes - Update congestion control for new connections without restarting the listener - Preserve existing connections and their selected congestion controller - Apply per-inbound log levels through the shared panel logger - Add lifecycle, authentication and TCP/UDP relay events without exposing secrets - Rate-limit repeated authentication and relay warnings - Support native and QUIC UDP relay modes on the same listener - Recover UDP associations after relay worker failures - Fix TCP relay cancellation, idle shutdown and half-close handling - Close active sessions when client credentials are revoked or disabled - Track traffic by immutable client statistics IDs across email and UUID changes - Prevent ambiguous accounting and duplicate UUIDs within TUIC inbounds - Persist pending traffic in a durable shutdown journal - Replay journal batches transactionally without duplicate accounting - Report server shutdown failures through the shared logger - Preserve legacy flat and nested TUIC settings compatibility - Normalize congestion controller values consistently across backend and frontend - Preserve controller, UDP mode and SNI in client links and subscriptions - Separate client profile options from server settings in the TUIC form - Keep certificate path autofill explicit when changing client SNI - Align UDP packet size validation with protocol limits - Simplify and localize TUIC field hints and certificate autofill messages - Add controller, TCP/UDP, logging and live settings update tests - Add accounting identity, journal replay and shutdown regression tests - Add relay recovery, session revocation and legacy frontend form tests * fix(service): alias the TUIC duplicate-UUID subquery for PostgreSQL < 16 syncInboundClients runs a COUNT(*) FROM (subquery) for every client sync, whatever the protocol. PostgreSQL before 16 rejects a FROM subquery with no alias, so on the distro PostgreSQL install.sh provisions (14 on Ubuntu 22.04, 15 on Debian 12) every client add or edit failed with SQLSTATE 42601. Reproduced against postgres:15 with the new env-gated test. * fix(database): create tuic_traffic_receipts through the model migration AddTuicTrafficBatch issued CREATE TABLE IF NOT EXISTS at runtime, a schema change outside db.go. The table was invisible to allModels and migrationModels, so x-ui migrate-db dropped the receipts and a retained journal could be counted twice after a SQLite to PostgreSQL move. It is now a GORM model in both lists, and the insert uses OnConflict DoNothing. * chore(tuic): skip the ICMP-dependent relay test on Windows, drop dead collectors Go disables SIO_UDP_CONNRESET on Windows, so a dead UDP bridge never fails a read there and TestAudit3UDPAssociationMustRecoverAfterBridgeReadFailure was red on every Windows run. Server.CollectTotalTraffic and Manager.CollectTraffic had no caller. * refactor(tuic): serve TUIC on apernet/quic-go instead of a personal fork The native server depended on github.com/poise52/quic-go, a personal fork of apernet/quic-go patched only to pick the congestion controller before the handshake. That put a second QUIC/TLS stack in the binary that no upstream security fix reaches. apernet/quic-go is already in the graph through xray-core and exposes SetCongestionControl, so BBR is now installed on each accepted connection with Xray's own congestion.UseBBR; the cross-module BBR adapter is gone. apernet ships New Reno as its only built-in sender, so a cubic setting is served as new_reno server-side (clients still get cubic in their profile). The test inspectors now read the sender under congestionMutex, which the post-handshake install writes under. Linux loopback, single stream through Xray: 2428 -> 3383 Mbit/s (bbr). --------- Co-authored-by: Sanaei <ho3ein.sanaei@gmail.com>
2026-10-02 17:56:02 -07:00
config.json | mtproto | mtproto/* | tuic | tuic/* | tuic-server | tuic-server-*) continue ;;
fix(install): preserve custom bin/ files (e.g. hand-added geoip) across updates (#6152) * fix(install): preserve custom bin/ files (e.g. hand-added geoip) across updates Every reinstall/update wipes /usr/local/x-ui/ wholesale and re-extracts the release tarball, which only ships known assets (xray/mtg binaries, the bundled geoip*/geosite*.dat sets). A user-reported real incident: a hand-placed custom geoip file referenced from a routing rule via "ext:<file>:<code>" got silently deleted on update, and Xray refused to start at all afterward ("failed to open <file>: no such file or directory"), taking down every inbound until the file was manually restored from the user's own backup. install_x-ui now backs up the old bin/ before the wipe and restores, after extraction, only the files the fresh release doesn't provide -- bundled assets still get the newer per-release copy, nothing custom silently disappears. Verified in isolation: standard files (geoip.dat, the xray binary) end up as the fresh release's copy; a custom file absent from the release survives untouched. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix: harden the bin/ snapshot-and-restore against the review round on #6152 - Replace the mktemp+cp snapshot with a same-filesystem mv of bin/ aside: an unchecked mktemp failure previously made the very next line copy bin/'s contents into "/" (empty custom_bin_backup + trailing slash), and a silently-ignored cp failure (stderr redirected, exit code never checked) could leave a truncated custom geo file that gets "restored" as if it were intact. A rename is atomic and needs no extra disk space, removing both failure modes at once; if it fails, back off cleanly and say so instead of proceeding as if a backup exists. - Add a trap so an interrupted update (Ctrl-C, signal) between the backup and the restore doesn't leave the snapshot (which contains bin/config.json and every mtproto client's FakeTLS secret) sitting around indefinitely; the two exit-path cleanups this replaces are gone since the trap now covers those exits too. - Move the restore below the arm arch-rename/chmod block instead of before it, so xray-linux-arm32/mtg-linux-arm already exist under their final names and don't get needlessly restored-then-overwritten and misreported as "custom". - Exclude bin/config.json and bin/mtproto/*.toml from the restore: those are the panel's own generated runtime state (internal/xray/process.go, internal/mtproto/manager.go), not admin-placed files, and restoring a stale one only resurrects dead state or recreates bin/mtproto/ with the wrong (more permissive) directory mode. - Match symlinks in the restore's find, not just plain files -- cp -a already preserves them in the snapshot, but the restore loop was silently dropping them, which is exactly the failure mode (a geo file symlinked in from elsewhere) this PR set out to fix. - Quote the two new xui_folder expansions. - Extend the non-interactive smoke test to reinstall over an existing install with a sentinel file in bin/, asserting it survives and that the bundled geoip.dat is still the release's own copy -- the update path this PR touches had no CI coverage at all before this. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
2026-08-14 07:41:45 -07:00
esac
if [[ ! -e "bin/${rel}" ]]; then
mkdir -p "bin/$(dirname "${rel}")"
cp -a "${f}" "bin/${rel}"
restored_custom_bin+=("${rel}")
fi
done < <(find "${custom_bin_backup}" \( -type f -o -type l \) -print0)
rm -rf "${custom_bin_backup}"
custom_bin_backup=""
if [[ ${#restored_custom_bin[@]} -gt 0 ]]; then
echo -e "${green}Restored custom file(s) in bin/ not shipped by this release: ${restored_custom_bin[*]}${plain}"
fi
fi
trap - EXIT INT TERM
feat(tuic): implement native in-process Go TUIC v5 server (#6577) * feat(tuic): implement native in-process Go TUIC v5 server - Implement native TUIC v5 protocol server on pure Go using quic-go - Bridge decrypted TCP/UDP traffic into Xray-core via loopback SOCKS5 inbound - Support full Xray routing rules (geosite/geoip) and cascading outbounds - Implement atomic per-client traffic accounting with TotalGB and ExpiryTime - Add automatic legacy cleanup for older Rust tuic-server binaries, configs, and orphaned processes - Eliminate external Rust tuic-server downloads from install/CI scripts * fix(tuic): address traffic accounting, client reload, and socket lifecycle issues * fix(service): update checkTuicSocksReverseConflict to use bindAddr for listenOverlaps * fix(tuic): resolve traffic double-accounting, UDP fragmentation, and socket lifecycle issues * feat(tuic): complete native Go integration and address audit findings - Integrate an isolated QUIC fork pinned to a specific commit - Preserve original QUIC dependencies for Xray, Hysteria and Gin - Apply BBR, CUBIC and Reno to server connections and exported client profiles - Bridge Xray BBR with correct monotonic time and congestion type conversions - Handle congestion sender recreation after PMTU changes - Update congestion control for new connections without restarting the listener - Preserve existing connections and their selected congestion controller - Apply per-inbound log levels through the shared panel logger - Add lifecycle, authentication and TCP/UDP relay events without exposing secrets - Rate-limit repeated authentication and relay warnings - Support native and QUIC UDP relay modes on the same listener - Recover UDP associations after relay worker failures - Fix TCP relay cancellation, idle shutdown and half-close handling - Close active sessions when client credentials are revoked or disabled - Track traffic by immutable client statistics IDs across email and UUID changes - Prevent ambiguous accounting and duplicate UUIDs within TUIC inbounds - Persist pending traffic in a durable shutdown journal - Replay journal batches transactionally without duplicate accounting - Report server shutdown failures through the shared logger - Preserve legacy flat and nested TUIC settings compatibility - Normalize congestion controller values consistently across backend and frontend - Preserve controller, UDP mode and SNI in client links and subscriptions - Separate client profile options from server settings in the TUIC form - Keep certificate path autofill explicit when changing client SNI - Align UDP packet size validation with protocol limits - Simplify and localize TUIC field hints and certificate autofill messages - Add controller, TCP/UDP, logging and live settings update tests - Add accounting identity, journal replay and shutdown regression tests - Add relay recovery, session revocation and legacy frontend form tests * fix(service): alias the TUIC duplicate-UUID subquery for PostgreSQL < 16 syncInboundClients runs a COUNT(*) FROM (subquery) for every client sync, whatever the protocol. PostgreSQL before 16 rejects a FROM subquery with no alias, so on the distro PostgreSQL install.sh provisions (14 on Ubuntu 22.04, 15 on Debian 12) every client add or edit failed with SQLSTATE 42601. Reproduced against postgres:15 with the new env-gated test. * fix(database): create tuic_traffic_receipts through the model migration AddTuicTrafficBatch issued CREATE TABLE IF NOT EXISTS at runtime, a schema change outside db.go. The table was invisible to allModels and migrationModels, so x-ui migrate-db dropped the receipts and a retained journal could be counted twice after a SQLite to PostgreSQL move. It is now a GORM model in both lists, and the insert uses OnConflict DoNothing. * chore(tuic): skip the ICMP-dependent relay test on Windows, drop dead collectors Go disables SIO_UDP_CONNRESET on Windows, so a dead UDP bridge never fails a read there and TestAudit3UDPAssociationMustRecoverAfterBridgeReadFailure was red on every Windows run. Server.CollectTotalTraffic and Manager.CollectTraffic had no caller. * refactor(tuic): serve TUIC on apernet/quic-go instead of a personal fork The native server depended on github.com/poise52/quic-go, a personal fork of apernet/quic-go patched only to pick the congestion controller before the handshake. That put a second QUIC/TLS stack in the binary that no upstream security fix reaches. apernet/quic-go is already in the graph through xray-core and exposes SetCongestionControl, so BBR is now installed on each accepted connection with Xray's own congestion.UseBBR; the cross-module BBR adapter is gone. apernet ships New Reno as its only built-in sender, so a cubic setting is served as new_reno server-side (clients still get cubic in their profile). The test inspectors now read the sender under congestionMutex, which the post-handshake install writes under. Linux loopback, single stream through Xray: 2428 -> 3383 Mbit/s (bbr). --------- Co-authored-by: Sanaei <ho3ein.sanaei@gmail.com>
2026-10-02 17:56:02 -07:00
rm -f bin/tuic-server bin/tuic-server-* > /dev/null 2>&1 || true
rm -rf bin/tuic > /dev/null 2>&1 || true
2026-05-19 04:09:35 -07:00
# Update x-ui cli and se set permission
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
mv -f "${xui_script_temp}" /usr/bin/x-ui
if [[ $? -ne 0 ]]; then
rm -f "${xui_script_temp}"
echo -e "${red}Failed to install x-ui.sh${plain}"
exit 1
fi
2026-05-19 04:09:35 -07:00
chmod +x /usr/bin/x-ui
mkdir -p /var/log/x-ui
config_after_install
2026-05-19 04:09:35 -07:00
# Etckeeper compatibility
if [ -d "/etc/.git" ]; then
if [ -f "/etc/.gitignore" ]; then
if ! grep -q "x-ui/x-ui.db" "/etc/.gitignore"; then
echo "" >> "/etc/.gitignore"
echo "x-ui/x-ui.db" >> "/etc/.gitignore"
echo -e "${green}Added x-ui.db to /etc/.gitignore for etckeeper${plain}"
fi
else
echo "x-ui/x-ui.db" > "/etc/.gitignore"
echo -e "${green}Created /etc/.gitignore and added x-ui.db for etckeeper${plain}"
fi
fi
2026-05-04 04:20:24 -07:00
if [[ $release == "alpine" ]]; then
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
xui_rc_temp="/etc/init.d/x-ui.tmp.$$"
rm -f "${xui_rc_temp}"
fix(install): fetch x-ui.sh and unit files from the installed release tag (#6391) * fix(install): fetch x-ui.sh and unit files from the installed release tag install.sh and update.sh pin the panel archive to a release tag but always took x-ui.sh, x-ui.rc and the service units from main, so the management script and the binary of one installation came from different commits: the fail2ban templates and setting flags the script writes drift silently against an older binary, two installs of the same tag differ, and a reviewed or digest-pinned installer still runs unreviewed code from main. Use the same ref as the archive, keeping main only for the rolling dev-latest build. The menu's "update menu" and update_shell paths now fetch the script matching the installed version and fall back to main with a visible notice when no script is published for it. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): fall back to main for files a pinned tag does not publish Review follow-ups. install.sh accepts tags down to v2.3.5, but x-ui.rc only exists from v2.8.4 and the split x-ui.service.* files are newer still, so pinning those to the tag made an Alpine install of an old tag 404 after the previous install was already removed. Probe the tag for each file and fall back to main with a notice when it is missing, as the menu already does for x-ui.sh. The fail2ban auto-setup probe also trusted the exit status of 'x-ui setup-fail2ban', but scripts before v3.4.0 have no such subcommand and exit 0 from the usage banner, so the installer reported a setup that never ran. Skip with a notice when the installed script does not know the subcommand. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): refuse a tag that does not publish a needed script Falling back to main reintroduced the binary/script mismatch the tag pinning exists to remove, and it fired at points where install.sh and update.sh have already stopped and removed the previous installation -- so the quiet path was also the one that could not be undone. Probe the tag instead, before anything is touched, for every file that is always fetched from GitHub (x-ui.sh, plus x-ui.rc on Alpine), and abort with the HTTP status when one is missing. The unit files stay unprobed: they are only fetched when the release tarball omits them, so an old tag that ships x-ui.service inside its tarball still installs. Their existing failure message now names the ref it tried. Also tighten the setup-fail2ban probe to the dispatcher's case arm rather than any mention of the string, which also matches a comment. --------- Co-authored-by: Claude <noreply@anthropic.com>
2026-09-03 07:50:53 -07:00
curl -fLRo "${xui_rc_temp}" "https://raw.githubusercontent.com/MHSanaei/3x-ui/${script_ref}/x-ui.rc"
2025-09-25 05:34:12 -07:00
if [[ $? -ne 0 ]]; then
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
rm -f "${xui_rc_temp}"
2026-05-19 04:09:35 -07:00
echo -e "${red}Failed to download x-ui.rc${plain}"
exit 1
fi
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if [[ ! -s "${xui_rc_temp}" ]]; then
rm -f "${xui_rc_temp}"
echo -e "${red}Downloaded x-ui.rc is empty${plain}"
exit 1
fi
mv -f "${xui_rc_temp}" /etc/init.d/x-ui
if [[ $? -ne 0 ]]; then
rm -f "${xui_rc_temp}"
echo -e "${red}Failed to install x-ui.rc${plain}"
exit 1
fi
2026-05-19 04:09:35 -07:00
chmod +x /etc/init.d/x-ui
rc-update add x-ui
rc-service x-ui start
else
# Install systemd service file
service_installed=false
if [ -f "x-ui.service" ]; then
echo -e "${green}Found x-ui.service in extracted files, installing...${plain}"
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if _install_xui_service_unit "x-ui.service" "false"; then
2026-05-19 04:09:35 -07:00
service_installed=true
2026-01-02 22:27:39 -08:00
fi
fi
2026-05-19 04:09:35 -07:00
if [ "$service_installed" = false ]; then
case "${release}" in
ubuntu | debian | armbian)
2026-01-02 22:27:39 -08:00
if [ -f "x-ui.service.debian" ]; then
2026-05-19 04:09:35 -07:00
echo -e "${green}Found x-ui.service.debian in extracted files, installing...${plain}"
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if _install_xui_service_unit "x-ui.service.debian" "false"; then
2026-01-02 22:27:39 -08:00
service_installed=true
fi
fi
2026-05-04 04:20:24 -07:00
;;
arch | manjaro | parch)
if [ -f "x-ui.service.arch" ]; then
2026-05-19 04:09:35 -07:00
echo -e "${green}Found x-ui.service.arch in extracted files, installing...${plain}"
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if _install_xui_service_unit "x-ui.service.arch" "false"; then
service_installed=true
fi
fi
2026-05-04 04:20:24 -07:00
;;
*)
2026-01-02 22:27:39 -08:00
if [ -f "x-ui.service.rhel" ]; then
2026-05-19 04:09:35 -07:00
echo -e "${green}Found x-ui.service.rhel in extracted files, installing...${plain}"
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if _install_xui_service_unit "x-ui.service.rhel" "false"; then
2026-01-02 22:27:39 -08:00
service_installed=true
fi
fi
2026-05-04 04:20:24 -07:00
;;
esac
2026-01-02 22:27:39 -08:00
fi
2026-05-19 04:09:35 -07:00
# If service file not found in tar.gz, download from GitHub
if [ "$service_installed" = false ]; then
echo -e "${yellow}Service files not found in tar.gz, downloading from GitHub...${plain}"
case "${release}" in
ubuntu | debian | armbian)
fix(install): fetch x-ui.sh and unit files from the installed release tag (#6391) * fix(install): fetch x-ui.sh and unit files from the installed release tag install.sh and update.sh pin the panel archive to a release tag but always took x-ui.sh, x-ui.rc and the service units from main, so the management script and the binary of one installation came from different commits: the fail2ban templates and setting flags the script writes drift silently against an older binary, two installs of the same tag differ, and a reviewed or digest-pinned installer still runs unreviewed code from main. Use the same ref as the archive, keeping main only for the rolling dev-latest build. The menu's "update menu" and update_shell paths now fetch the script matching the installed version and fall back to main with a visible notice when no script is published for it. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): fall back to main for files a pinned tag does not publish Review follow-ups. install.sh accepts tags down to v2.3.5, but x-ui.rc only exists from v2.8.4 and the split x-ui.service.* files are newer still, so pinning those to the tag made an Alpine install of an old tag 404 after the previous install was already removed. Probe the tag for each file and fall back to main with a notice when it is missing, as the menu already does for x-ui.sh. The fail2ban auto-setup probe also trusted the exit status of 'x-ui setup-fail2ban', but scripts before v3.4.0 have no such subcommand and exit 0 from the usage banner, so the installer reported a setup that never ran. Skip with a notice when the installed script does not know the subcommand. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): refuse a tag that does not publish a needed script Falling back to main reintroduced the binary/script mismatch the tag pinning exists to remove, and it fired at points where install.sh and update.sh have already stopped and removed the previous installation -- so the quiet path was also the one that could not be undone. Probe the tag instead, before anything is touched, for every file that is always fetched from GitHub (x-ui.sh, plus x-ui.rc on Alpine), and abort with the HTTP status when one is missing. The unit files stay unprobed: they are only fetched when the release tarball omits them, so an old tag that ships x-ui.service inside its tarball still installs. Their existing failure message now names the ref it tried. Also tighten the setup-fail2ban probe to the dispatcher's case arm rather than any mention of the string, which also matches a comment. --------- Co-authored-by: Claude <noreply@anthropic.com>
2026-09-03 07:50:53 -07:00
service_unit_url="https://raw.githubusercontent.com/MHSanaei/3x-ui/${script_ref}/x-ui.service.debian"
2026-05-19 04:09:35 -07:00
;;
arch | manjaro | parch)
fix(install): fetch x-ui.sh and unit files from the installed release tag (#6391) * fix(install): fetch x-ui.sh and unit files from the installed release tag install.sh and update.sh pin the panel archive to a release tag but always took x-ui.sh, x-ui.rc and the service units from main, so the management script and the binary of one installation came from different commits: the fail2ban templates and setting flags the script writes drift silently against an older binary, two installs of the same tag differ, and a reviewed or digest-pinned installer still runs unreviewed code from main. Use the same ref as the archive, keeping main only for the rolling dev-latest build. The menu's "update menu" and update_shell paths now fetch the script matching the installed version and fall back to main with a visible notice when no script is published for it. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): fall back to main for files a pinned tag does not publish Review follow-ups. install.sh accepts tags down to v2.3.5, but x-ui.rc only exists from v2.8.4 and the split x-ui.service.* files are newer still, so pinning those to the tag made an Alpine install of an old tag 404 after the previous install was already removed. Probe the tag for each file and fall back to main with a notice when it is missing, as the menu already does for x-ui.sh. The fail2ban auto-setup probe also trusted the exit status of 'x-ui setup-fail2ban', but scripts before v3.4.0 have no such subcommand and exit 0 from the usage banner, so the installer reported a setup that never ran. Skip with a notice when the installed script does not know the subcommand. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): refuse a tag that does not publish a needed script Falling back to main reintroduced the binary/script mismatch the tag pinning exists to remove, and it fired at points where install.sh and update.sh have already stopped and removed the previous installation -- so the quiet path was also the one that could not be undone. Probe the tag instead, before anything is touched, for every file that is always fetched from GitHub (x-ui.sh, plus x-ui.rc on Alpine), and abort with the HTTP status when one is missing. The unit files stay unprobed: they are only fetched when the release tarball omits them, so an old tag that ships x-ui.service inside its tarball still installs. Their existing failure message now names the ref it tried. Also tighten the setup-fail2ban probe to the dispatcher's case arm rather than any mention of the string, which also matches a comment. --------- Co-authored-by: Claude <noreply@anthropic.com>
2026-09-03 07:50:53 -07:00
service_unit_url="https://raw.githubusercontent.com/MHSanaei/3x-ui/${script_ref}/x-ui.service.arch"
2026-05-19 04:09:35 -07:00
;;
*)
fix(install): fetch x-ui.sh and unit files from the installed release tag (#6391) * fix(install): fetch x-ui.sh and unit files from the installed release tag install.sh and update.sh pin the panel archive to a release tag but always took x-ui.sh, x-ui.rc and the service units from main, so the management script and the binary of one installation came from different commits: the fail2ban templates and setting flags the script writes drift silently against an older binary, two installs of the same tag differ, and a reviewed or digest-pinned installer still runs unreviewed code from main. Use the same ref as the archive, keeping main only for the rolling dev-latest build. The menu's "update menu" and update_shell paths now fetch the script matching the installed version and fall back to main with a visible notice when no script is published for it. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): fall back to main for files a pinned tag does not publish Review follow-ups. install.sh accepts tags down to v2.3.5, but x-ui.rc only exists from v2.8.4 and the split x-ui.service.* files are newer still, so pinning those to the tag made an Alpine install of an old tag 404 after the previous install was already removed. Probe the tag for each file and fall back to main with a notice when it is missing, as the menu already does for x-ui.sh. The fail2ban auto-setup probe also trusted the exit status of 'x-ui setup-fail2ban', but scripts before v3.4.0 have no such subcommand and exit 0 from the usage banner, so the installer reported a setup that never ran. Skip with a notice when the installed script does not know the subcommand. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): refuse a tag that does not publish a needed script Falling back to main reintroduced the binary/script mismatch the tag pinning exists to remove, and it fired at points where install.sh and update.sh have already stopped and removed the previous installation -- so the quiet path was also the one that could not be undone. Probe the tag instead, before anything is touched, for every file that is always fetched from GitHub (x-ui.sh, plus x-ui.rc on Alpine), and abort with the HTTP status when one is missing. The unit files stay unprobed: they are only fetched when the release tarball omits them, so an old tag that ships x-ui.service inside its tarball still installs. Their existing failure message now names the ref it tried. Also tighten the setup-fail2ban probe to the dispatcher's case arm rather than any mention of the string, which also matches a comment. --------- Co-authored-by: Claude <noreply@anthropic.com>
2026-09-03 07:50:53 -07:00
service_unit_url="https://raw.githubusercontent.com/MHSanaei/3x-ui/${script_ref}/x-ui.service.rhel"
2026-05-19 04:09:35 -07:00
;;
esac
fix: make all self-managed file downloads/installs atomic, with real completion status (#5711) * fix(script): download the live x-ui.sh script atomically before replacing it update_menu(), update_shell(), and update.sh's update_x-ui() all overwrote /usr/bin/x-ui in place via `curl -o`, truncating and rewriting the same inode a currently-running x-ui process may still be reading from. A network hiccup or slow write during that overwrite leaves a half-old/half-new script on disk, which then fails with bogus syntax errors on the next run. Download to /usr/bin/x-ui-temp and `mv -f` into place instead, matching the atomic pattern install.sh already uses. Also fixes update_menu() checking chmod's exit code instead of curl's, which meant a failed download could still report "Update successful." * fix(script): close remaining gaps in the atomic script-update path Code review of the previous commit found the atomic mv fix was itself incomplete: - None of the mv -f calls checked their exit status, so a failed move fell through to chmod and "success" messaging while /usr/bin/x-ui stayed on the old file. - update_shell()'s `[[ -s x-ui-temp ]]` guard couldn't tell "curl -z got a 304, nothing to do" from "a stale temp file survived an earlier crashed run" -- the latter could get moved into place with no freshness check. - update_menu(), update_shell(), and update_x-ui() all hardcoded the same /usr/bin/x-ui-temp path, so two concurrent updates (e.g. a cron auto-update racing an interactive menu update) could collide. - update.sh's update_x-ui() was missing the non-empty-file guard update_shell() already had. x-ui.sh's update_menu() and update_shell() now share a replace_xui_script() helper that uses a PID-suffixed temp path (/usr/bin/x-ui-temp.$$), pre-cleans it before every attempt, and checks the exit status of curl, the non-empty test, and mv before treating the update as successful. update.sh's update_x-ui() gets the same sequence inlined (it's fetched as a standalone script and can't call x-ui.sh's function), closing the missing-guard gap and using its own unique temp path. * fix(script,panel): harden the remaining self-update download paths install.sh had the same unguarded /usr/bin/x-ui-temp overwrite the two already-fixed scripts had: no exit-status check on mv, and a fixed temp name shared with x-ui.sh/update.sh's (now-unique) temp files. Give it its own PID-suffixed temp path, an empty-file guard, and an mv exit-status check, matching the pattern used there. Audited the web dashboard's Go-native updater (panel.go) for the same bug class: it already uses os.CreateTemp for a genuinely unique temp file and cleans up via both a deferred Remove and a shell EXIT trap, so it was never exposed to the fixed-path race. It was missing a check for a zero-byte download (a 200 OK with an empty body would chmod +x and exec an empty script) -- added that alongside the existing size cap. Not addressed here: once startUpdate()'s child process starts, the Go service releases it and returns success immediately. If update.sh fails partway through, the still-running old panel keeps answering /status, so the frontend's poll can report success with no update having happened. Fixing that needs update.sh to signal completion status back and the frontend to check it -- a separate follow-up. * feat(panel): report real completion status for the web self-update Fixes the fire-and-forget gap flagged in the atomic-overwrite fix: once startUpdate() launches update.sh detached, the Go service had no way to learn whether it actually succeeded. If update.sh failed partway (network drop, disk full, permission denied), the still-running old panel kept answering /status, so the frontend's poll reported success with nothing having changed. update.sh now writes its outcome to a small JSON status file (/etc/x-ui/update-status.json by default) via `trap ... EXIT`, which covers every exit path in the script -- including the two bare `exit 1` call sites that don't go through the existing _fail() helper. The Go service generates a run ID before launching, passes it and the status path to update.sh via XUI_UPDATE_RUN_ID/XUI_UPDATE_STATUS_FILE, and a new GET /panel/api/server/getUpdateStatus endpoint reports it back. The frontend now polls that instead of blindly trusting HTTP reachability, and shows a distinct error or "couldn't confirm" message instead of silently reloading into a false success. Adversarial review of this surfaced three more issues, fixed here: - No lock stopped two concurrent /updatePanel calls from launching two update.sh runs that would race each other on the actual update work (tar extraction, service unit swap). Added an in-memory guard with a 5-minute self-expiring window, so a run that never reaches a terminal state doesn't lock out retries indefinitely. - XUI_UPDATE_RUN_ID is read from the environment and was interpolated unquoted into the status JSON; a malformed value would produce invalid JSON. Now validated as digits-only before use. - The run ID is a UnixNano timestamp (19 digits), sent as a raw JSON number it would lose precision in JavaScript (past Number.MAX_SAFE_INTEGER), letting two different runs round to the same value on the wire and defeat the whole comparison. It's now a decimal string end to end (Go, the status file, and the generated frontend type). install.sh's equivalent temp-file/mv path and the Go-native downloadPanelUpdater() path were audited for the same bug classes during this work; findings from that audit were addressed separately. * fix(panel): release the update lock as soon as the run finishes An exhaustive multi-angle review of the whole branch (12 finder angles, 3-vote adversarial verification, a fresh-eyes sweep) surfaced a real bug in the concurrency guard added in the previous commit, plus several smaller issues; this fixes what's actionable now. The bug: acquireUpdateSlot only ever released on the 5-minute stale timeout or if launching itself failed. If update.sh launched fine but failed fast (bad GitHub API response, "x-ui not installed", any of its early exit paths), the status file correctly reported "failed" within seconds, but a retry was still rejected with "a panel update is already in progress" for up to 5 more minutes -- the guard never looked at the very status file this branch built to know a run was done. It now tracks which run ID currently holds the slot and checks that run's own status before falling back to the timeout, so a fast failure clears the way for an immediate retry. Added a regression test for this, plus one confirming a stale, unrelated runID can't be mistaken for the current run finishing. Also: - Added a genuinely concurrent test for the guard: 200 goroutines racing acquireUpdateSlot, asserting exactly one wins. The previous tests only ever called it from one goroutine, so they gave no signal if the mutex's check-then-set were silently broken -- verified this by temporarily removing the lock and confirming the old tests still passed while the new one caught it immediately under -race. - Removed the redundant upfront "pending" status write: GetUpdateStatus already defaults a missing/stale file to pending, and the frontend matches by run ID regardless, so the write changed no observable behavior. Deleted writeUpdateStatus entirely since that was its only caller. - Renamed replace_xui_script()'s unclear "conditional" parameter to use_if_modified_since, matching what it actually controls. - Added HTTP-level tests for the new getUpdateStatus endpoint, including a regression test that the runId wire format is a JSON string (decoding into a Go string field fails outright if it were ever a bare number). updatePanel's actual launch path is not covered: on a Linux test runner it would make a real network call and could exec a real update.sh, so only its non-Linux guard path is safely testable without mocking. Not fixed here, tracked separately: the same unsafe-overwrite pattern this branch eliminated for /usr/bin/x-ui is still present for the systemd unit file install in update.sh and install.sh (lower severity since systemd only reads it on daemon-reload, not continuously); and startUpdate's systemd-run-vs-detached-fallback branching has no test coverage since testing it safely needs dependency injection this fix doesn't warrant bundling in. * fix(script): make systemd unit file installation atomic Same anti-pattern as the /usr/bin/x-ui overwrite fixed earlier: every site that lands the systemd unit at ${xui_service}/x-ui.service -- copying it from the extracted release tarball, or falling back to a GitHub download per distro family -- wrote straight onto the live path via cp/curl, no temp file, no verification. A network drop mid-download or an interrupted cp leaves the unit file truncated; systemd then fails to parse it on the next daemon-reload/start, leaving the panel unable to come up until an operator manually re-copies a good unit file. Lower severity than the /usr/bin/x-ui case (systemd only reads this file on demand at daemon-reload time, not continuously the way bash interprets a running script line by line), but it's the identical gap, just left uncovered when that fix landed. Added a small shared helper in both update.sh and install.sh -- _install_xui_service_unit() -- covering both source types (cp from the tarball, curl from GitHub): write to a PID-suffixed temp file, verify the copy/download succeeded and the result is non-empty, then mv -f into place and check that exit status too, matching the pattern already used for /usr/bin/x-ui. All 4 cp sites and the 3-way curl fallback in each file now go through it; verified no other site writes new content to the unit path (the remaining ${xui_service} references are a pre-install existence check, an rm during old-version cleanup, and the chown/chmod that already ran after the file is safely in place -- none of those need atomicity). Verified with bash -n on both files, plus a standalone scratch test exercising cp-success, cp-with-missing-source, cp-with-empty-source, and curl-failure paths: on every failure the previous, good unit file content is left untouched and no temp file is leaked behind. * fix(script): make Alpine's OpenRC init script install atomic; drop a stray comment A final maximum-rigor review of the whole PR (12 finder angles including a repo-wide sweep for any remaining instance of the bug class this PR fixes) found two more real issues: - Alpine's /etc/init.d/x-ui startup script is downloaded via a bare `curl -fLRo` straight onto the live path in both update.sh and install.sh -- the exact same unguarded-overwrite pattern already fixed for /usr/bin/x-ui and the systemd unit file, just left uncovered on the OpenRC side. A network drop mid-download truncates the live init script; OpenRC then fails to source/execute it on the next start, leaving the panel unable to come up. Fixed with the same temp-file + non-empty check + mv -f (with its own exit-status check) pattern used everywhere else in this PR. Verified with bash -n and a standalone scratch-script test covering success, empty-download, and destination-preserved-on-failure paths. - internal/web/service/panel/panel_test.go had one line-level `//` comment on a call site, which the root CLAUDE.md's hard rule ("No // line comments in committed Go/TS... rename instead of annotating") explicitly prohibits. The comment duplicated context already stated in the test's own doc comment two lines above, so it's simply removed rather than reworded. Also flagged, deliberately not bundled here since it's a different subsystem: x-ui.sh's update_geofiles() downloads Xray's live geoip.dat/geosite.dat with the same unguarded curl -o pattern. Tracked as its own follow-up. * fix(script): make geo-data file downloads atomic Same anti-pattern as /usr/bin/x-ui, the systemd unit file, and the Alpine init script fixed in prior PRs: update_geofiles() downloaded Xray's live geoip.dat/geosite.dat (and the IR/RU variants) with curl writing straight onto the exact path Xray reads at runtime (internal/xray/process.go's GetGeoipPath/GetGeositePath), no temp file, no verification. The existing check only inspected the reported HTTP status via -w '%{http_code}', not file integrity, so a network drop mid-download could leave a truncated .dat file on disk that passes the status check. Xray then fails to parse it on the next restart/reload, breaking any routing rules that reference geoip:/ geosite:. The -z conditional-GET usage needed care here: the original code pointed both -z and -o at the same live path. Fixed by pointing -z at the live file (to keep the "already current" freshness check) while -o writes to a PID-suffixed temp file, matching the pattern already proven in x-ui.sh's replace_xui_script(). Verified with a local HTTP server that a 304 response leaves the temp file untouched/nonexistent (so the existing "already up to date" branch still works unchanged), and added a non-empty check plus a checked mv -f before treating a download as installed. Verified with bash -n and an end-to-end scratch test against a local server covering: fresh download, 304-not-modified, empty response body, and a 404 -- confirming a failure at any stage leaves the previous good .dat file completely untouched and no temp file behind. * fix(script): verify the release tarball extraction, not just the download The final maximum-rigor review found the most significant remaining gap in this whole effort: update.sh and install.sh check the tarball download's exit status, but never check tar's exit status, and never verify the extracted x-ui binary actually exists before continuing. Worse, by the time extraction runs, the previous installation has already been stopped and deleted -- there's no rollback. A truncated download that still passes curl's own check, or a tar failure (disk full, killed process), left the panel silently in a broken half-state: chmod/config/service-install all continued to run against a missing or empty binary, with no error surfaced anywhere. This is the same bug class as everything else in this PR (unverified write to a path something then depends on), just for the tarball itself rather than a single file -- and it also covers the geo-data files this PR already fixed once for the interactive/cron path, since they ship inside this same tarball on every panel update. Added: a non-empty check on the downloaded archive (both files, both install.sh call sites) and a check that tar succeeded and produced a non-empty x-ui binary before proceeding, failing loudly with a message that explicitly says the previous install is already gone, since silently continuing here is worse than anywhere else in this PR. This doesn't make the multi-file extraction fully atomic (that would mean extracting to a temp directory and atomically swapping the whole install tree into place, a materially larger restructuring than anything else in this PR) -- but it closes the "fails silently, user discovers it days later when Xray can't start" gap, which was the actual reported problem this whole effort traces back to. Also fixed, all much smaller: - replace_xui_script() in x-ui.sh implicitly returned chmod's exit status instead of success, so a successful atomic install could be reported as failed if chmod transiently failed after the mv already landed the new script. Added an explicit `return 0`. - update_geofiles() had no default case branch; an unrecognized argument would silently reuse whatever dat_files/dat_source values a previous call left in the un-scoped globals instead of failing. Currently unreachable (all three call sites pass fixed literals) but cheap, defensive, and worth having. - internal/web/controller/server.go's updatePanel has one branch (an unparseable "dev" form value) that's both untested and safe to test on any platform, since it's rejected before any real exec/network call. Added the missing test case. Verified: bash -n on all three scripts; an empirical scratch test covering an empty downloaded archive, a corrupt (non-gzip) archive, and a successfully-extracting-but-empty archive, confirming each is caught before the script proceeds; full go build/vet/test -race across the whole module; frontend generation confirmed still in sync. * fix(panel): base the update-slot staleness fallback on process liveness Addresses the automated review on the upstream PR (MHSanaei/3x-ui#5711). Blocking finding: acquireUpdateSlot's staleness fallback freed the update slot purely on elapsed wall-clock time (5 minutes), with no check on whether the update.sh process it launched was actually still running. update.sh runs install_base() (apt-get/dnf/pacman update and install) before update_x-ui even starts, plus several GitHub downloads (release tarball, x-ui.sh, and possibly a service unit or x-ui.rc) -- on a slow or throttled host, a small VPS being the typical deployment target for this project, that alone can plausibly exceed 5 minutes with nothing wrong. A second /updatePanel call arriving in that window (an admin retrying after the frontend's 90s poll times out, or overlapping master-node bulk-update calls) would launch a second update.sh, racing the exact rm/tar/mv/systemctl sequence this whole PR exists to make safe. Fixed by recording the launched process's PID (detached-fallback path only; the systemd-run path's own process has already exited by the time startUpdate returns, so it never learns update.sh's real PID) and checking it via the standard POSIX kill(pid, 0) liveness probe before treating a run as stale, following the existing panel_unix.go / panel_other.go platform-split pattern already used for setDetachedProcess. A confirmed-alive process now keeps the slot held past updateStaleAfter (raised from 5 to 20 minutes as a safer baseline for the systemd-run path, which still has no way to check liveness directly). updateHardCeiling (2 hours) is an absolute backstop so a genuinely wedged run can never lock out retries permanently even on the PID-tracked path. Added two regression tests exercising the new logic (gated to Linux, since processAlive is a no-op stub elsewhere): a live PID keeps the slot held past the stale window, and the hard ceiling overrides liveness. Traced both by hand against the new acquireUpdateSlot logic; could not execute-verify processAlive itself on this Windows dev machine (no WSL distro installed, and installing one felt disproportionate to validate kill(pid, 0), an extremely well-established POSIX primitive), but cross-compiled clean for linux/amd64 and this repo's CI runs the real test suite on Linux. Also fixed, both suggestions from the same review: - install.sh: two failure paths right after tarball extraction were exiting without cleaning up the already-downloaded x-ui.sh temp file (xui_script_temp), leaving it behind. Every other new failure branch in this PR removes its temp file before exiting; these two now do too. - frontend/src/pages/api-docs/endpoints.ts: updatePanel's doc entry did not reflect that a successful response now carries an obj with runId. Added an inline response example matching the existing pattern used for other ad hoc (non-schema-backed) responses like getWebCertFiles. Verified: go build/vet clean on both windows (native) and a linux/amd64 cross-compile; full go test ./... clean; go test -race on the panel and controller packages; bash -n on all three shell scripts; npm run gen confirms the openapi.json diff is exactly the new response example with no stray changes to src/generated; TestAPIRoutesDocumented still passes.
2026-07-02 09:19:33 -07:00
if ! _install_xui_service_unit "$service_unit_url" "true"; then
fix(install): fetch x-ui.sh and unit files from the installed release tag (#6391) * fix(install): fetch x-ui.sh and unit files from the installed release tag install.sh and update.sh pin the panel archive to a release tag but always took x-ui.sh, x-ui.rc and the service units from main, so the management script and the binary of one installation came from different commits: the fail2ban templates and setting flags the script writes drift silently against an older binary, two installs of the same tag differ, and a reviewed or digest-pinned installer still runs unreviewed code from main. Use the same ref as the archive, keeping main only for the rolling dev-latest build. The menu's "update menu" and update_shell paths now fetch the script matching the installed version and fall back to main with a visible notice when no script is published for it. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): fall back to main for files a pinned tag does not publish Review follow-ups. install.sh accepts tags down to v2.3.5, but x-ui.rc only exists from v2.8.4 and the split x-ui.service.* files are newer still, so pinning those to the tag made an Alpine install of an old tag 404 after the previous install was already removed. Probe the tag for each file and fall back to main with a notice when it is missing, as the menu already does for x-ui.sh. The fail2ban auto-setup probe also trusted the exit status of 'x-ui setup-fail2ban', but scripts before v3.4.0 have no such subcommand and exit 0 from the usage banner, so the installer reported a setup that never ran. Skip with a notice when the installed script does not know the subcommand. Assisted-by: Claude Code:claude-fable-5-1 * fix(install): refuse a tag that does not publish a needed script Falling back to main reintroduced the binary/script mismatch the tag pinning exists to remove, and it fired at points where install.sh and update.sh have already stopped and removed the previous installation -- so the quiet path was also the one that could not be undone. Probe the tag instead, before anything is touched, for every file that is always fetched from GitHub (x-ui.sh, plus x-ui.rc on Alpine), and abort with the HTTP status when one is missing. The unit files stay unprobed: they are only fetched when the release tarball omits them, so an old tag that ships x-ui.service inside its tarball still installs. Their existing failure message now names the ref it tried. Also tighten the setup-fail2ban probe to the dispatcher's case arm rather than any mention of the string, which also matches a comment. --------- Co-authored-by: Claude <noreply@anthropic.com>
2026-09-03 07:50:53 -07:00
echo -e "${red}Failed to install x-ui.service from GitHub (${script_ref}) -- the release tarball did not ship one either${plain}"
2026-05-19 04:09:35 -07:00
exit 1
fi
service_installed=true
fi
if [ "$service_installed" = true ]; then
echo -e "${green}Setting up systemd unit...${plain}"
chown root:root ${xui_service}/x-ui.service > /dev/null 2>&1
chmod 644 ${xui_service}/x-ui.service > /dev/null 2>&1
systemctl daemon-reload
systemctl enable x-ui
systemctl start x-ui
else
echo -e "${red}Failed to install x-ui.service file${plain}"
exit 1
fi
fi
2026-05-04 04:20:24 -07:00
# IP Limit relies on fail2ban; install + configure it now so the feature
# works out of the box (no-op when XUI_ENABLE_FAIL2BAN=false). Never fatal.
setup_fail2ban
2026-05-19 04:09:35 -07:00
echo -e "${green}x-ui ${tag_version}${plain} installation finished, it is running now..."
2023-02-09 11:18:06 -08:00
echo -e ""
echo -e "┌───────────────────────────────────────────────────────┐
│ ${blue}x-ui control menu usages (subcommands):${plain} │
│ │
│ ${blue}x-ui${plain} - Admin Management Script │
│ ${blue}x-ui start${plain} - Start │
│ ${blue}x-ui stop${plain} - Stop │
│ ${blue}x-ui restart${plain} - Restart │
│ ${blue}x-ui status${plain} - Current Status │
│ ${blue}x-ui settings${plain} - Current Settings │
│ ${blue}x-ui enable${plain} - Enable Autostart on OS Startup │
│ ${blue}x-ui disable${plain} - Disable Autostart on OS Startup │
│ ${blue}x-ui log${plain} - Check logs │
│ ${blue}x-ui banlog${plain} - Check Fail2ban ban logs │
│ ${blue}x-ui update${plain} - Update │
│ ${blue}x-ui legacy${plain} - Legacy version │
│ ${blue}x-ui install${plain} - Install │
│ ${blue}x-ui uninstall${plain} - Uninstall │
2025-12-03 15:05:21 -08:00
└───────────────────────────────────────────────────────┘"
2023-02-09 11:18:06 -08:00
}
echo -e "${green}Running...${plain}"
install_base
2026-05-19 04:09:35 -07:00
install_x-ui $1