Move DATABASE_URL, ADMIN_TOKEN, ARTDAG_CLUSTER_KEY to .env
and use env_file on all services. This means docker stack deploy
no longer needs env vars sourced in the shell, and repeat deploys
won't trigger spurious restarts on unchanged services.
GPU worker gets its own .env.gpu with fully resolved cross-VPC URLs.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
docker stack deploy does not read .env files automatically
(unlike docker compose), so ${VAR} substitutions resolve to
empty strings. Source .env to export vars before deploying.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Rsync was leaving dirty working trees on the deploy server, causing
git pull conflicts during manual deploys. Use git fetch + reset instead.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Remove gpu-worker.yml (no GPU server)
- Fix ci.yml: install ssh/rsync in job container, remove GPU steps
- Remove source mounts from l1-server and l1-worker so they use image code
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Central config with logging on startup
- Hot reload support for GPU worker (docker-compose.gpu-dev.yml)
- Quick deploy script (scripts/gpu-dev-deploy.sh)
- GPU/CPU frame compatibility tests
- CI/CD pipeline for GPU worker (.gitea/workflows/gpu-worker.yml)
- Standardize GPU_PERSIST default to 0 across all modules
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
