Haven

Add QEMU and Buildx steps to your workflow, then specify platforms in the build-push action. Use this template:

build-image:
  runs-on: ubuntu-latest
  steps:
    - name: Checkout
      uses: actions/checkout@v4
    - name: Set up QEMU
      uses: docker/setup-qemu-action@v3
    - name: Set up Docker Buildx
      uses: docker/setup-buildx-action@v3
    - name: Login to GitHub Container Registry
      uses: docker/login-action@v3
      with:
        registry: ghcr.io
        username: ${{ github.repository_owner }}
        password: ${{ secrets.GITHUB_TOKEN }}
    - name: Build and push image
      uses: docker/build-push-action@v6
      with:
        context: .
        platforms: linux/amd64,linux/arm64
        push: true
        tags: your-tag

Replace your-tag and adjust registry/credentials as needed. Add needs: create-tag if you have a tagging step.

The PostgreSQL database "haven" already exists message is normally harmless. The restart is caused by a stale Puma PID file, while missing images/customizations are usually caused by the storage path changing from /storage to /app/storage.

1. Edit docker-compose.yml and update the Haven volume mount:

volumes:
  - ./storage:/app/storage

Replace any previous ./storage:/storage mount.

2. Recreate the container and remove the stale PID file:

docker compose down
docker compose rm -f haven
docker compose up -d

If the PID error continues, remove it directly while the container is briefly running:

docker exec -it haven-haven-1 rm -f /app/tmp/pids/server.pid
docker compose restart haven

As a temporary workaround, pin the previous working image:

image: ghcr.io/havenweb/haven:e18a854