Backup & Continuity

You Have a Backup — Does It Restore?

A nightly job backs up the databases and cluster configuration, a restore drill proves the backup actually comes back, and the nightly watchdog checks that it is fresh.

Fail-fast on critical sources

If the databases can't be backed up, the job isn't quietly counted as a success — it exits with a critical error.

Proven restore

The drill restores the backup into isolated, throwaway containers; its effect on the live system is zero.

What's Inside

What Do You Actually Have When Disaster Strikes?

Continuity starts not with taking a backup but with knowing you can come back from it.

Nightly Automatic Backup

PostgreSQL, the panel identity/audit database and Slurm accounting are backed up every night; retention is configurable.

Image Registry

The OCI image registry can be included in the backup — your node images never have to be rebuilt from scratch.

Config Bundle

Inventory and licensing configuration export as a checksummed JSON bundle with secrets redacted.

Mandatory Dry Run

Importing a bundle produces a dry-run report first; nothing is applied before you have seen what would change.

Nightly Watchdog

Deployment drift, unexpected server errors and backup freshness are checked every night; findings go into the panel's event stream.

Support Bundle

Open a support case with a one-click diagnostic bundle — no need to log into the server to collect logs.

Not a Backup — a Guarantee of Return

Taking a backup is easy; the real question is whether it comes back when disaster strikes.

The drill spins up throwaway containers on a network entirely separate from production, using temporary storage; backup files are read-only and everything is cleaned up in every case.

Checksum consistency, that the database dumps genuinely restore, that the schema version is at the expected level, and that panel users can be read back from the restored copy.

The bundle is SHA-256 checksummed and contains no secrets; it can be hand-carried into a closed network safely.

Long-running backup and restore jobs run as background operations whose progress is followed from the panel.

Want to see it live on your cluster? Get in touch!
%