Skip to content

Backups

The backups list is served from a local cache, kept fresh by a background sync job. If the cache has gone stale, opening the Backups page triggers a refresh automatically, so a freshly connected host doesn’t show a permanently empty list waiting on a sync nobody kicked off.

You can restore a backup in place (overwriting the original guest) or as a new guest alongside it. An in-place restore is refused in any of these cases, checked in this order:

  • the original guest no longer exists,
  • the restore target is the same guest the backup itself came from,
  • you haven’t confirmed the restore, or
  • the guest is currently running.

The running-guest refusal is the one that matters most: restoring in place over a running guest would destroy its current disk state mid-operation. Stop the guest first, or restore as a new guest instead if you want to keep the original running while you inspect the backup.

Before pruning old backups, you can preview exactly what would be deleted: a read-only, dry-run pass that can’t delete anything by itself. Run the real prune only once the preview looks right.