Make container more robust (#95)
Reviewed-on: #95 12 files changed (+149, -66) Key changes: 1. Monolithic container — Replaced the two-service compose setup (backend + frontend/caddy) with a single timetracker container. Caddy is now built into the image rather than running as a separate container. 2. Supervisord process manager — Added supervisor.conf and installed supervisor in the Dockerfile. entrypoint.sh now delegates to supervisord to manage three processes: Caddy, Gunicorn, and Qcluster — replacing manual trap/signaling logic. 3. Bundled Caddy — The Dockerfile now downloads and installs Caddy v2.9.1 directly into the image (/usr/local/bin/caddy). The Caddyfile was updated to use reverse_proxy localhost:8001 and serves static files from /home/timetracker/app/static. 4. Configurable deployment — Added .env.example with configurable environment variables: TZ, PUID/PGID, TIMETRACKER_EXTERNAL_PORT, DATA_DIR, CSRF_TRUSTED_ORIGINS. docker-compose.yml now references these with sensible defaults. 5. UID/GID flexibility — entrypoint.sh uses usermod/groupmod at startup to remap the timetracker user to the host-specified PUID/PGID, avoiding permission issues with mounted volumes. 6. Database & static files — settings.py now respects DATA_DIR env var for the SQLite database path. STATIC_ROOT changed to BASE_DIR / "static". 7. Dev improvements — New Caddyfile.dev (with browse enabled for static files) and updated Makefile dev-prod target runs Caddy alongside Django in development. 8. Tests — Re-enabled the test step in the Docker build GitHub Actions workflow.
This commit was merged in pull request #95.
This commit is contained in:
@@ -41,9 +41,10 @@ caddy:
|
||||
|
||||
dev-prod: migrate collectstatic
|
||||
@npx concurrently \
|
||||
--names "Django,Django-Q" \
|
||||
"PROD=1 uv run python -m gunicorn --bind 0.0.0.0:8001 timetracker.asgi:application -k uvicorn.workers.UvicornWorker"
|
||||
"uv run manage.py qcluster"
|
||||
--names "Caddy,Django,Django-Q" \
|
||||
"caddy run --config Caddyfile.dev" \
|
||||
"PROD=1 uv run python -m gunicorn --bind 0.0.0.0:8001 timetracker.asgi:application -k uvicorn.workers.UvicornWorker" \
|
||||
"PROD=1 uv run manage.py qcluster"
|
||||
|
||||
dumpgames:
|
||||
uv run python manage.py dumpdata --format yaml games --output tracker_fixture.yaml
|
||||
|
||||
Reference in New Issue
Block a user