cardinal 1.23 — a persistent restart supervisor and scheduled backups

Containers recover on boot, crash loops get a budget, and backups are scheduled and verified.

1.23 is about keeping containers alive and restorable:

  • Persistent systemd supervisoralways and unless-stopped containers

restart automatically after reboot or crash, with configurable

--restart-delay.

  • Crash-loop protection--restart-max-attempts / --restart-window

put a budget on rapid restarts; once exhausted the container is blocked

(restart_blocked) until an explicit cardinal start.

  • Scheduled backupsbackup enable/disable/status with configurable

intervals and retention; backup verify checks an archive's checksum.

  • Offline image verificationcardinal verify IMAGE checks layer

digests against the locally stored manifest.

  • Runtime hardening — zombie-exit detection, rm tombstones that stop

supervisor races, safe OCI extraction, and protected bind sources.

The complete bilingual (EN/RU) documentation suite ships with it.