Files
Reactive-Resume/compose.yml
T
Amruth Pillai bdfb854602 fix: resolve storage healthcheck path via LOCAL_STORAGE_PATH env var (#3004)
* fix: resolve local data directory to /app/data in production Docker

In the official Docker image, cwd is /app/apps/web (set via WORKDIR), but
the data volume is mounted at /app/data. Without pnpm-workspace.yaml present
in the runtime image, findWorkspaceRoot() returns null, so getLocalDataDirectory()
fell back to <cwd>/data = /app/apps/web/data, which the node user has no
permission to create. This caused the storage healthcheck to fail with
EACCES.

Add a production fallback: when cwd ends in apps/web, resolve the data
directory to two levels up (matching /app/data in the official image).

Re-resolves #2990.

https://claude.ai/code/session_015pSTtukxf7mFTty2Y6PHZf

* fix: replace apps/web heuristic with LOCAL_STORAGE_PATH env var

The previous fix special-cased a cwd ending in apps/web to land on /app/data,
but the heuristic could false-positive on any path with that suffix and was
fragile to Dockerfile changes. pnpm-workspace.yaml is never copied into the
runtime image, so the workspace-root walk was also dead code in production.

Replace the heuristic with an explicit LOCAL_STORAGE_PATH env var:
- Set LOCAL_STORAGE_PATH=/app/data in the Dockerfile (single source of truth).
- Add LOCAL_STORAGE_PATH to the env schema; storage and statistics services
  pass it through to getLocalDataDirectory.
- getLocalDataDirectory now uses the override when set, else workspace root
  (dev), else cwd/data.
- New Nitro plugin validates the resolved local data directory at startup
  and refuses to boot with a clear error if it isn't writable, surfacing
  permission issues immediately instead of at first upload/healthcheck.
- Document the new variable in .env.example and the Docker self-hosting docs.

https://claude.ai/code/session_015pSTtukxf7mFTty2Y6PHZf

* fix: address review feedback on storage path handling

- apps/web/plugins/2.storage.ts: use the default-import style for
  node:fs/promises (matches the rest of the repo, sidesteps any
  named-export concerns for fs.constants).
- packages/env/src/server.ts: reject relative LOCAL_STORAGE_PATH values
  via a zod refinement. Relative paths would be resolved against cwd,
  which differs between dev and Docker — exactly the same surprise the
  original bug had. Failing fast at config validation time gives a
  clear error before the server boots.

https://claude.ai/code/session_015pSTtukxf7mFTty2Y6PHZf

* fix: update data volume configuration in Docker Compose and enhance Nitro plugin

* fix: remove "Can I customize the templates?" FAQ entry from multiple language files

---------

Co-authored-by: Claude <noreply@anthropic.com>
2026-05-08 16:05:47 +02:00

99 lines
2.4 KiB
YAML

name: reactive_resume
services:
postgres:
image: postgres:latest
restart: unless-stopped
networks:
- data_network
environment:
- POSTGRES_DB=postgres
- POSTGRES_USER=postgres
- POSTGRES_PASSWORD=postgres
ports:
- "5432:5432"
volumes:
- postgres_data:/var/lib/postgresql
healthcheck:
test: ["CMD", "pg_isready", "-U", "postgres", "-d", "postgres"]
start_period: 10s
interval: 30s
timeout: 10s
retries: 3
seaweedfs:
image: chrislusf/seaweedfs:latest
restart: unless-stopped
command: server -s3 -filer -dir=/data -ip=0.0.0.0
networks:
- storage_network
environment:
- AWS_ACCESS_KEY_ID=seaweedfs
- AWS_SECRET_ACCESS_KEY=seaweedfs
ports:
- "8333:8333"
volumes:
- seaweedfs_data:/data
healthcheck:
test: ["CMD", "wget", "-q", "-O", "/dev/null", "http://localhost:8888"]
start_period: 10s
interval: 30s
timeout: 10s
retries: 3
seaweedfs_create_bucket:
image: quay.io/minio/mc:latest
restart: on-failure
entrypoint: >
/bin/sh -c "
sleep 5;
mc alias set seaweedfs http://seaweedfs:8333 seaweedfs seaweedfs;
mc mb seaweedfs/reactive-resume;
exit 0;
"
networks:
- storage_network
depends_on:
seaweedfs:
condition: service_healthy
reactive_resume:
build:
context: .
dockerfile: Dockerfile
networks:
- data_network
- storage_network
ports:
- "3000:3000"
environment:
- APP_URL=http://localhost:3000
- DATABASE_URL=postgresql://postgres:postgres@postgres:5432/postgres
- AUTH_SECRET=change-me-to-a-secure-secret-key-in-production
- S3_ACCESS_KEY_ID=seaweedfs
- S3_SECRET_ACCESS_KEY=seaweedfs
- S3_ENDPOINT=http://seaweedfs:8333
- S3_BUCKET=reactive-resume
- S3_FORCE_PATH_STYLE=true
volumes:
- ./data:/app/data
depends_on:
postgres:
condition: service_healthy
seaweedfs_create_bucket:
condition: service_completed_successfully
healthcheck:
test: ["CMD", "node", "-e", "fetch('http://127.0.0.1:3000/api/health').then((r) => { if (!r.ok) process.exit(1); }).catch(() => process.exit(1));"]
start_period: 10s
interval: 30s
timeout: 10s
retries: 3
volumes:
postgres_data:
seaweedfs_data:
networks:
data_network:
storage_network: