After a series of disks going bad, I hastily created a new config and while my 3 cache pools (vm, incoming, docker_appdata) were functional but with the GUI showing unmountable I tried fixing that. After a restart, while my empty pool did get recognized correctly (the one i used as a poc) and without having touched the other disks I can no longer mount them or use them in the corresponding pool with the data accessible.

Not sure how to better describe it.

tldr: my nvme disk for my docker_appdata cache pool is no longer recognized by unraid due to no filesystem present but the data is still on the disk.

unraid-diagnostics-20260919-1600.zip

Edited September 19Sep 19 by hardw1red

Please post the output from:

blkid

zpool import

btrfs fi show

In the atttached file

blkid.txt

The diagnostics show no recognized filesystem signatures on the appdata NVMe or either incoming-pool device. ZFS also cannot discover the appdata pool. The labels may have been erased or damaged, although we should check the partition layout before concluding that.

Please post the output of these read-only commands:

zdb -l /dev/nvme0n1p1

zdb -l /dev/sdg1

zdb -l /dev/sde1

fdisk -l /dev/nvme0n1 /dev/sdg /dev/sde

These device names match the diagnostics. If you have rebooted since then, please post fresh diagnostics first because the names can change.

Do not format or recreate either pool. If the partition boundaries are correct and no usable labels remain, there is unlikely to be a straightforward recovery.

In the attached file

zdb_devnvme0n1p1.txt

I have the CommunityApplicationsDataBackup with a 2 week old backup which is not ideal but it'll do in case this is unfixable.

None of the four ZFS labels can be decoded on any of those three partitions, so a normal pool import looks unlikely.

The partition listings show the current layout, but we cannot confirm whether it matches the original layout. This means we cannot yet conclude that all data is lost.

If you can accept losing the changes since your two-week-old backup, restoring it looks like the most practical option. First, verify that the backup contains the appdata you need.

Do you also need to recover anything from the incoming pool?

If you need the newer or unbacked-up data, do not format or recreate either pool. Please post any diagnostics from before this happened and describe exactly what you changed when attempting the original fix. Preserve the affected devices for imaging before further recovery attempts.

Appreciate the assistance; I went ahead and recreated the pool and restored from my backup. Good opportunity to fine-tune the AppdataBackup plugin as well.