Major fixes for production deployment issues:
1. Migration System:
- Add safe migration runner that handles existing schema
- Create migration helper functions for idempotent operations
- Auto-detect existing tables and mark migrations as applied
- Handle "relation already exists" errors gracefully
2. Production Initialization:
- Create init-production.sh script for proper startup sequence
- Fix directory creation and permissions
- Add admin user creation from environment variables
- Ensure proper service initialization order
3. Documentation:
- Add comprehensive PRODUCTION_DEPLOYMENT_GUIDE.md
- Add MIGRATION_ERROR_FIX.md for immediate issue resolution
- Document all known production issues and solutions
- Include backup/restore procedures
4. Safety Improvements:
- Add migrate:safe npm script for production use
- Update wait-for-db.sh to use safe migrations in production
- Add proper error handling and logging
This resolves the "relation already exists" error and prevents similar
issues in future deployments. The safe migration system can handle both
fresh installations and existing databases.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fix database connection error "getaddrinfo ENOTFOUND postgres"
- Add wait-for-db.sh script to ensure PostgreSQL is ready before starting
- Fix email processor initialization timing issue
- Add missing storage path environment variables
- Add database dependency to backend service
- Enhance health check endpoint with database connectivity check
- Update production database defaults to match docker-compose
- Install postgresql-client in Docker image for health checks
- Document all required environment variables in .env.example
Fixes immediate production deployment failures and ensures proper service startup order.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
paul