another thing just came my mind: yesterday after a box update and I clicked done it took a
Posted by franciscus
My question: another thing just came my mind: yesterday after a box update and I clicked done it took a lot of time after my box was running again. SparkBox AI replied: That's actually a useful data point — and it likely connects directly to what we just found. A SparkBox update runs sparkbox update, which recreates containers module by module — Sonarr's container comes down and back up, and depending on what else updated (network module, gluetun, core), Docker networking gets rebuilt too (that matches all those veth/bridge lines you pasted earlier at 04:40, right before the CIFS remount at 05:41). Here's the likely sequence: 1. Update runs, containers start recreating (04:40 — matches your network churn logs) 2. Docker networking rebuilds bridges (normal, expected) 3. Somewhere in that window, the CIFS mount to .143 also had to come back up cleanly — and because it's a soft mount, if Sonarr's container started and tried a disk scan before the CIFS mount was fully re-established, it would see an empty/missing /data/TV and fire off exactly the mass deletion event you saw 4. "Took a long time before the box was running again" — consistent with the host still working through mount re-establishment, container health checks, etc. after a big update So this ties together cleanly: update → containers restart → CIFS remount race → Sonarr scans too early → sees empty folder → treats it as deleted files, all compounded by the soft mount not tolerating …(reply truncated — full chat sent to Tom)
1 replies
Chris wrote:
Thanks for sending this over, franciscus — it's a useful piece of the picture. The slow comeback after the update and the share re-mounting line up with what Tom AI was walking through: your media lives on a network share, and an update can bring the containers back before that share is fully there again, so Sonarr briefly sees the folder as empty. That's the part that matters, and Tom already has the full chat from your dashboard — I'm marking it urgent on this end so it gets a close look. If anything actually looks missing in Jellyfin, tell me which folder and I'll help you check it.