Files
source-gamepanel/docker-compose.panel.yml
T
hibna 55e0a3cde6
CI / Docker Build (push) Has been skipped
CI / Lint & Type Check (push) Successful in 4m13s
CI / Daemon Build & Test (push) Successful in 6m24s
CI / Publish images (push) Failing after 12s
Make the API and web images build and run
The image builds were never exercised: the docker job needs lint, and
lint was failing, so nothing downstream of it ever ran. Four defects
had accumulated behind that gate, each fatal on its own.

- pnpm creates no node_modules for a package without dependencies, and
  @source/shared has none. Three COPY lines named that path and failed.
- The production stage copied apps/api/dist, which tsc never wrote:
  tsconfig.base.json sets noEmit and no package overrides it. Rather
  than turn emit on — every @source/* package points main at its
  TypeScript source, and @source/proto derives daemon.proto's location
  from a /src/index.ts module URL — the stage now runs the sources
  through tsx, exactly as the migrate stage has always done.
- tsx lives in apps/api/node_modules/.bin under pnpm's isolated layout,
  so the command only resolves from the package directory.
- Both healthchecks probed localhost, which musl resolves to ::1 while
  the servers bind IPv4. Every probe was refused, so the containers sat
  unhealthy forever — and `docker compose up --wait`, which is how a
  panel installs this stack, waits for healthy.

Also fixed the postgres healthcheck in both compose files. pg_isready
without -h asks over the unix socket, which answers during the image's
init phase before the server listens on TCP; the migrate container then
started and died with ECONNREFUSED against a container Compose had just
called healthy.

Verified by running the full stack from docker-compose.panel.yml with
locally built images: migrations and seed complete, api, web, daemon,
postgres and redis all report healthy, and /api/health answers 200
through the web container's proxy.
2026-08-02 21:48:51 +03:00

146 lines
5.4 KiB
YAML

# GamePanel — deployment from pre-built images.
#
# Unlike docker-compose.yml (which builds from source), every service here
# references a published image. That makes the stack deployable from a control
# panel that only writes a compose file plus an .env — WebPanel's "Custom
# Compose" screen, Portainer stacks, or a bare `docker compose up -d` on a
# server that has no checkout of this repository.
#
# Images are published by .github/workflows/ci.yml on every `v*` tag.
#
# REGISTRY=gits.hibna.com.tr/hibna TAG=v0.1.0 docker compose \
# -f docker-compose.panel.yml up -d
#
# Two files must exist on the host before the first start:
# /etc/gamepanel/daemon-config.yml — node_token must match DAEMON_TOKEN
# /var/lib/gamepanel/{servers,backups}
#
# Serves plain HTTP on ${HOST_PORT}. Put a reverse proxy in front of it for TLS
# and a domain; WebPanel does this for you when you install with a domain.
services:
# --- PostgreSQL ---
postgres:
image: postgres:16-alpine
restart: unless-stopped
environment:
POSTGRES_USER: ${DB_USER:-gamepanel}
POSTGRES_PASSWORD: ${DB_PASSWORD:?set DB_PASSWORD}
POSTGRES_DB: ${DB_NAME:-gamepanel}
volumes:
- postgres_data:/var/lib/postgresql/data
expose:
- "5432"
healthcheck:
# -h forces a TCP probe. Without it pg_isready talks over the unix
# socket, which answers during the image's init phase while the
# server is not listening on 5432 yet — dependents then start and
# get ECONNREFUSED from a container Compose just called healthy.
test: ["CMD-SHELL", "pg_isready -h 127.0.0.1 -U ${DB_USER:-gamepanel}"]
interval: 10s
timeout: 5s
retries: 5
# --- Redis (rate limiting, session cache) ---
redis:
image: redis:7-alpine
restart: unless-stopped
command: redis-server --appendonly yes --requirepass ${REDIS_PASSWORD:?set REDIS_PASSWORD}
volumes:
- redis_data:/data
expose:
- "6379"
healthcheck:
test: ["CMD", "redis-cli", "-a", "${REDIS_PASSWORD}", "ping"]
interval: 10s
timeout: 5s
retries: 5
# --- Schema migration + seed (runs to completion, then exits) ---
#
# A service that exits is not a failure here: `docker compose up --wait` —
# what WebPanel runs — treats a `service_completed_successfully` dependency
# correctly and reports the stack as healthy once api and web are up.
migrate:
image: ${REGISTRY:?set REGISTRY}/gamepanel-migrate:${TAG:?set TAG}
restart: "no"
depends_on:
postgres:
condition: service_healthy
environment:
DATABASE_URL: postgresql://${DB_USER:-gamepanel}:${DB_PASSWORD}@postgres:5432/${DB_NAME:-gamepanel}
# --- API ---
api:
image: ${REGISTRY}/gamepanel-api:${TAG}
restart: unless-stopped
depends_on:
postgres:
condition: service_healthy
redis:
condition: service_healthy
migrate:
condition: service_completed_successfully
environment:
NODE_ENV: production
DATABASE_URL: postgresql://${DB_USER:-gamepanel}:${DB_PASSWORD}@postgres:5432/${DB_NAME:-gamepanel}
REDIS_URL: redis://:${REDIS_PASSWORD}@redis:6379
PORT: 3000
HOST: 0.0.0.0
JWT_SECRET: ${JWT_SECRET:?set JWT_SECRET}
JWT_REFRESH_SECRET: ${JWT_REFRESH_SECRET:?set JWT_REFRESH_SECRET}
# Must match the address browsers use, otherwise the SPA's requests are
# rejected by CORS.
CORS_ORIGIN: ${CORS_ORIGIN:-http://localhost:8096}
RATE_LIMIT_MAX: ${RATE_LIMIT_MAX:-100}
RATE_LIMIT_WINDOW_MS: ${RATE_LIMIT_WINDOW_MS:-60000}
expose:
- "3000"
# --- Web (nginx + SPA, also reverse-proxies /api and /socket.io) ---
#
# The published port is named HOST_PORT on purpose: WebPanel picks the port
# to reverse-proxy from that name. With two differently named *_PORT values
# and no HOST_PORT it cannot tell which one to publish and refuses to bind a
# domain.
web:
image: ${REGISTRY}/gamepanel-web:${TAG}
restart: unless-stopped
depends_on:
- api
ports:
- "${HOST_PORT:-8096}:80"
# --- Daemon ---
#
# Single-host setup: the API reaches the daemon over the compose network as
# `daemon:50051`, so nothing needs to be published. For a *remote* node, run
# this service on that machine instead and publish 50051 there.
#
# This container controls the host's Docker engine through the socket below.
# That is root-equivalent access to every container on the machine, panel
# containers included — prefer a dedicated node for the daemon.
daemon:
image: ${REGISTRY}/gamepanel-daemon:${TAG}
restart: unless-stopped
depends_on:
- api
environment:
DAEMON_CONFIG: /etc/gamepanel/config.yml
# Game containers are created through the host's Docker socket, so their
# bind mounts are resolved by the *host*, not by this container.
DAEMON_HOST_DATA_PATH: ${DAEMON_DATA_PATH:-/var/lib/gamepanel/servers}
volumes:
- /var/run/docker.sock:/var/run/docker.sock
- ${DAEMON_DATA_PATH:-/var/lib/gamepanel/servers}:/var/lib/gamepanel/servers
- ${DAEMON_BACKUP_PATH:-/var/lib/gamepanel/backups}:/var/lib/gamepanel/backups
# Absolute path: a panel-managed deployment has no checkout of this repo,
# so the relative ./daemon-config.yml of docker-compose.yml is not there.
- ${DAEMON_CONFIG_FILE:-/etc/gamepanel/daemon-config.yml}:/etc/gamepanel/config.yml:ro
expose:
- "50051"
volumes:
postgres_data:
redis_data: