Troubleshooting
Common issues with the agent and backups, and how to resolve them.
The agent shows as offline
A server is “offline” when it hasn’t checked in recently. On the server:
- Confirm the service is running:
systemctl status rested. Start it withsudo systemctl start rested. - Check the agent can reach Rested (outbound HTTPS). Firewalls or lost network are the usual cause.
- If the server was revoked or re-installed, generate a new enrollment token from its page and re-run the install command.
A backup failed
Open the repository to see the failure and a log excerpt. Common causes:
- Permission denied — the agent service runs as root, but mounted filesystems, container permissions, ACLs, or a protected restore target can still block access. Check the configured paths and the repository logs before changing permissions.
- Disk full — free space on the server, or on the restore target when restoring.
- Database connection failed — for database recipes, make sure the database or container is running and the configured host, port, and credentials are correct.
- Wrong credentials — a storage or database credential changed. Update it in the repository configuration.
restic is not installed
The agent downloads and pins a verified restic binary during install. If a backup reports restic missing, re-run the installer or sudo rested ensure-restic on the server, then retry.
The snapshot list is empty
Snapshots appear after the first successful backup and after the agent reports them back. If the list looks empty:
- Confirm at least one backup has completed successfully in the repository's history.
- Press Refresh on the Snapshots tab while the agent is online, so it can list the repository and report what it finds.
- If runs are failing before any snapshot is written, resolve the backup failure above first — there is nothing to list until one succeeds.
A backup is unexpectedly large or slow
restic deduplicates data, so size and duration usually settle after the first run:
- The first backup uploads a full copy of the selected paths. Later runs only send changed chunks, so they are smaller and faster.
- Exclude caches, build output, and temporary directories you do not need to recover — they inflate every backup.
- A run that keeps growing usually means large files are changing between backups (logs, database data directories). Back up a database dump instead of its live data files where you can.
Still stuck?
For recovery-specific questions — including restoring without the original machine — see Recovery & your encryption key.