github junkerderprovinz/bombvault v7.8.0

latest releases: v10.6.3, v10.6.2, v10.6.1...
3 months ago

BombVault 7.8.0 makes backup errors visible and fixable straight from the dashboard, fixes saving the VM backup order, and fixes several cross-instance restore issues, including one that could scatter a multi-folder container's appdata into the wrong place and one that left the restore destination folder unreadable and wrongly owned.

✨ Added

  • See and clear backup errors from the dashboard. The error count is now clickable and opens a panel that lists every failed backup run with its time, domain, target name and error message. Identical errors that fired across many targets (for example "backup path not mounted" hitting every container after a reboot) are grouped into a single row with a count, so a systemic problem is obvious at a glance instead of scrolling past dozens of identical rows. You can filter the list the same way as the Activity Log, and acknowledge errors right there, one at a time or with "Mark all resolved", without ever touching the database. Acknowledged errors drop off the count.

🐛 Fixed

  • Saving the VM backup order no longer fails with "Invalid name" for VMs whose names contain spaces. Names like "Windows 11" or "Home Assistant" were rejected by the stricter container-name check. The VM backup order now uses the VM-aware validator, matching the rest of the VM handling.
  • The cross-instance restore destination field no longer suggests a path that does not exist. The "Appdata destination" field showed a hardcoded user/bombvault/container example that read like a real chosen default but pointed at a path that is not there. It now shows the conventional appdata location user/appdata (that is, /mnt/user/appdata), and the VM disk field shows user/domains, so the greyed-out hint matches what a blank field actually uses.
  • Cross-instance restore no longer breaks containers with more than one appdata folder. A container that binds several separate appdata subfolders (for example SnapOtter/conf and SnapOtter/data) had each folder mapped independently by its own name, so a remapped restore dropped conf and data directly into the destination's appdata root instead of nesting them under the container's own folder. They landed next to, and could collide with, other containers' folders of the same name. Restores now keep the container-relative folder structure intact; containers with a single appdata folder (the common case) are unaffected. This also fixes a false "already contains data" prompt the old mapping caused when restoring such a container back in place.
  • The "already contains data" and similar restore-destination messages now show the real path. They were rendering the literal placeholder [path] instead of the folder in question, telling you nothing about what was actually in the way.
  • A remapped cross-instance restore destination folder is no longer left unreadable and wrongly owned. restic creates the destination folder for a remapped restore itself but never restores that folder's own owner or mode, only the files inside it. The folder is now created readable up front, and after a successful restore its original owner (for example nobody) and permissions are read back from the snapshot and reapplied. This applies to both container and VM restores, so a container or VM that needs a specific owner on its appdata or disk root to start correctly is no longer left broken after a cross-instance restore.

Upgrading

Just pull the new image.

Don't miss a new bombvault release

NewReleases is sending notifications on new releases.