Back to Changelog
v1.0.259
Migration reliability improvements: 5 live-test bug fixes and 5 long-run resilience enhancements
Improvements
5- SSH reconnect counter uses a 5-minute sliding window instead of a lifetime counter, preventing false lockouts during long migrations
- Migration checkpoint manager is now active: each completed site is checkpointed so restarts can resume from where they left off
- Backend restart now recovers active migrations to paused state, preserving on-disk checkpoints for operator resume
- Pre-migration disk space preflight check added with 20% safety margin to catch insufficient space before any data is transferred
- Pipeline step panics are now caught and reported as step errors, preventing entire migration runs from crashing on unexpected exceptions
Bug Fixes
5- PHP-FPM health check no longer spams logs for non-existent systemd units (LoadState not-found/masked silent skip)
- Plesk email discovery now uses direct SQL (psa.mail) as primary method, falling back to CLI and directory scan
- rndc reload failure now triggers rndc reconfig + retry before falling back to full reload
- WordPress wp-config.php is no longer flagged as modified when database credentials are already correct
- MariaDB 10.10+ UCA-1400 collations (utf8mb4_uca1400_ai_ci) are now converted to MySQL 8 compatible collations, preventing import failures