Blame my ADHD brain
Sorry
- Now we have the server working properly in automation using BullMQ and Redis
- It also presents an API endpoint that is used to accept connections for websockets by agents (WS or WSS)
- Updated the docker-compose.yml and its documentation
- Modified 20251005000000_add_user_sessions to check if table exists first
- Added existence checks for all indexes and foreign keys
- Migration now works for both fresh installs and 1.2.7 upgrades
- Prevents P3018 error by gracefully handling existing table
- Added comprehensive logging for debugging
- Simplified logic to focus on core issue: table exists but no migration record
- Creates migration record when user_sessions table exists from 1.2.7
- Prevents P3018 error by marking migration as already applied
- More reliable approach for production upgrades
- Resolve migration reconciliation conflicts
- Include updated migration that handles 1.2.7 upgrade scenario
- Merge automation features and Docker support
- Add case for table exists but no migration record (1.2.7 upgrade)
- Creates migration record for existing user_sessions table
- Prevents P3018 error when table exists from 1.2.7 installation
- Handles all upgrade scenarios properly
- Creates migration 20251004999999_reconcile_user_sessions_migration
- Runs before 20251005000000_add_user_sessions migration
- Properly handles failed migrations by marking them as rolled back first
- Handles migration name conflicts from 1.2.7 (add_user_sessions) to 1.2.8+ (20251005000000_add_user_sessions)
- Fixes failed migration states from upgrade attempts
- Works naturally within Prisma migration system
- No changes to Docker entrypoint or setup scripts needed
Resolved conflicts:
- backend/src/server.js: Kept automation routes alongside gethomepage routes
- frontend/src/pages/Queue.jsx: Kept deleted (replaced by Automation.jsx)
- setup.sh: Kept newer version date (2025-10-11)
This merge brings in all commits from main including:
- GetHomepage integration
- Version 1.2.9 updates
- Migration file renames
- Bug fixes and improvements
- Added gethomepageRoutes.js for GetHomepage integration
- Updated all package.json files to version 1.2.9
- Updated agent script to version 1.2.9
- Updated version fallbacks in versionRoutes.js and updateScheduler.js
- Updated setup.sh with version 1.2.9
- Merged GetHomepage integration UI (Integrations.jsx)
- Updated docker-entrypoint.sh from main
- Updated VersionUpdateTab component
- Combined automation and gethomepage routes in server.js
- Maintains both BullMQ automation and GetHomepage functionality
Added Indexing to the database for faster and efficient searching
Fixed some filtering from the hosts page relating to packages that need updating
Added buy me a coffee link (sorry and thank you <3)
Added profile session management
Added "Remember me" to bypass TFA using device fingerprint
Fixed profile name not being persistent after logout and login
- Update updateScheduler to use default GitHub repository
- Remove 'No GitHub repository configured' warning message
- Use same default fallback logic as version routes
- Add commit difference tracking between main branch and release tag
- Show how many commits main branch is ahead of current release
- Update UI to display branch status with clear messaging
- Fix linting issues with useCallback and unused parameters
- Simplify version display with My Version | Latest Release layout
The migration was dropping the constraint but not the underlying unique index.
In PostgreSQL, unique constraints and unique indexes can exist independently.
This caused auto-enrollment to fail with 'unique constraint violated' errors.
Added explicit DROP INDEX statement to ensure the unique index is removed,
allowing duplicate friendly_name values while machine_id remains unique.
- Add machine_id support for manual host creation from GUI
- Generate temporary 'pending-{uuid}' machine_id for new hosts
- Agent now collects and sends machine_id on every update
- Backend replaces pending machine_id with real one on first agent connection
- Remove unnecessary duplicate name check (friendly_name can be duplicated)
- Add get_machine_id() function to agent (reads from /etc/machine-id, /var/lib/dbus/machine-id, or generates fallback)
- Display IP address in Network tab on host details page
- Fix network tab visibility conditions to include host.ip
This ensures proper host identification using machine_id while maintaining backwards compatibility with API credentials as the primary authentication method.
- Added machine_id field to host details page
- Backend now returns machine_id in all host queries
- Users can search hosts by machine_id
- Added hostname index to schema for better performance
- Auto-enrollment endpoints now require and validate machine_id
- Check for duplicates by machine_id instead of friendly_name
- Added /hosts/check-machine-id endpoint for agent installer
- Bulk enrollment updated to handle machine_id
- Multiple hosts with same hostname now supported
- Add machine_id field to hosts schema (unique, indexed)
- Remove unique constraint from friendly_name (allow duplicate hostnames)
- Agent installer now generates/reads persistent machine_id
- Proxmox script retrieves machine_id from LXC containers
- Backend will check machine_id instead of hostname for duplicates
This allows multiple hosts with same hostname to coexist in PatchMon
