acodeaday

Appearance

Troubleshooting ​

Common issues and solutions when deploying or running acodeaday.

Database Issues ​

Connection failed / timeout ​

Symptoms: Backend fails to start, "connection refused", or timeout errors.

Solutions:

  1. Verify connection string format - Must use postgresql+asyncpg:// (not postgresql://)

  2. For Supabase Cloud - Use the Transaction pooler URL (port 6543), not direct connection:

    bash
    # ✅ Correct - Transaction pooler (port 6543)
postgresql+asyncpg://postgres.[ref]:[password]@aws-0-[region].pooler.supabase.com:6543/postgres

# ❌ Wrong - Direct connection (port 5432)
postgresql+asyncpg://postgres.[ref]:[password]@aws-0-[region].pooler.supabase.com:5432/postgres

    To get the correct string: Supabase Dashboard → Settings → Database → Connection string → URI → Transaction pooler

  3. For local Supabase with Docker - Use host.docker.internal instead of localhost:

    bash
    DATABASE_URL=postgresql+asyncpg://postgres:postgres@host.docker.internal:54322/postgres

  4. Check IP allowlist - In Supabase Dashboard → Settings → Database → Network, ensure your server's IP is allowed.

Authentication Issues ​

"Email not confirmed" error ​

Problem: User can't log in because email isn't confirmed.

Solutions:

  1. Check email inbox (including spam folder) for confirmation link

  2. Manually confirm in Supabase Dashboard:

    • Go to Authentication → Users
    • Find the user and click "Confirm email"

  3. Use service role key for auto-confirm (recommended):

    • Set SUPABASE_SERVICE_ROLE_KEY in your environment
    • The backend will auto-confirm users when creating them
    • Get this from Supabase Dashboard → Settings → API → service_role key

"Invalid credentials" error ​

Problem: Login fails even with correct email/password.

Solutions:

  1. Verify SUPABASE_URL and SUPABASE_KEY are correct
  2. Check Supabase project is active (not paused)
  3. Verify user exists in Supabase Dashboard → Authentication → Users
  4. Ensure email is confirmed (see above)

Frontend Issues ​

Browser showing old API URLs after config change ​

Problem: After changing VITE_API_URL or other environment variables, the browser still makes requests to the old URL.

Cause: Browser caches JavaScript files. The frontend replaces environment variable placeholders at container startup, but browsers serve cached versions.

Solutions:

  1. Hard refresh: Ctrl+Shift+R (Windows/Linux) or Cmd+Shift+R (Mac)

  2. Clear browser cache:

    • Chrome: Settings → Privacy → Clear browsing data → Cached images and files
    • Or: DevTools (F12) → Network tab → Check "Disable cache" → Refresh

  3. Try incognito/private mode to verify with a clean cache

  4. Log out and log back in to force fresh page load

After Changing Environment Variables

  1. Restart the frontend container
  2. Tell users to hard refresh their browsers

Frontend can't reach backend ​

Problem: API requests fail, CORS errors, or "Network Error".

Solutions:

  1. Verify VITE_API_URL is set correctly:

    • Local dev: http://localhost:8000
    • Subdomain routing: https://api.yourdomain.com
    • Path-based routing: /api or https://yourdomain.com/api

  2. Check CORS configuration - Backend's CORS_ORIGINS must include the frontend domain:

    bash
    CORS_ORIGINS=https://yourdomain.com,https://www.yourdomain.com

  3. Verify backend is running:

    bash
    curl https://api.yourdomain.com/health

Double /api/api in URLs (path-based routing) ​

Problem: Requests go to /api/api/... instead of /api/....

Cause: VITE_API_URL includes /api suffix AND frontend code adds /api prefix to endpoints.

Solutions:

  1. Use subdomain routing instead (recommended):

    • Backend at api.yourdomain.com
    • Set VITE_API_URL=https://api.yourdomain.com
    • No path stripping needed

  2. For path-based routing: Ensure the reverse proxy isn't stripping /api prefix, or adjust VITE_API_URL accordingly.

Judge0 Issues ​

Judge0 not starting ​

Problem: Judge0 container fails to start or stays unhealthy.

Cause: Judge0 requires privileged mode for sandboxed code execution.

Solution: Ensure privileged: true is set:

yaml
services:
  judge0-server:
    privileged: true  # Required!

WARNING

Some platforms (like certain Kubernetes setups) don't support privileged containers. Use an external Judge0 instance or RapidAPI in that case.

Code execution times out ​

Problem: Submissions hang or timeout.

Solutions:

  1. Check Judge0 is healthy:

    bash
    curl http://localhost:2358/about

  2. Check Judge0 workers are running:

    bash
    docker compose logs judge0-workers

  3. Restart Judge0 services:

    bash
    docker compose restart judge0-server judge0-workers

Deployment Issues (Coolify/VPS) ​

502 Bad Gateway ​

Problem: Reverse proxy returns 502 error.

Solutions:

  1. Check container is running:

    bash
    docker compose ps

  2. Check container logs:

    bash
    docker compose logs backend
docker compose logs frontend

  3. Verify port configuration in your reverse proxy matches the container's exposed port

SSL/HTTPS issues ​

Problem: Mixed content errors or SSL certificate issues.

Solutions:

  1. Ensure all URLs use HTTPS:

    • VITE_API_URL=https://... (not http)
    • VITE_SUPABASE_URL=https://...

  2. Verify SSL certificate is valid and not expired

  3. Check reverse proxy SSL configuration

Getting Help ​

If you're still stuck:

  1. Check container logs for detailed error messages:

    bash
    docker compose logs backend
docker compose logs frontend
docker compose logs judge0-server

  2. Search existing issues on GitHub

  3. Open a new issue with:

    • What you're trying to do
    • Error messages (from logs and browser console)
    • Your environment (Docker, Coolify, etc.)
    • Relevant config (sanitize secrets!)