OK I've got a problem here that I'm having some trouble resolving. Had a hard shutdown of my system and after reboot my ZFS array (12 disk RAIDZ2 single VDEV with mirrored set of Special devices and a mirrored set of log devices) is not usable any more.

The pool and special subpools were showing as one device in the pool but no devices listed. I tried to recreate the array as best I could by changing the numbers and putting all the devices back in, but it would not start saying that the subpools were 'orphans'. Did a "New Config" as I was running out of ideas and was able to put the array back together and start the array, but on starting the array it comes up as "Unmountable" and only allows me to format.

Here's the other weird part; when I go to stop the array all of a sudden it mounts the array and boom all the data is there but obviously all the services are shut down.

After a clean reboot, I can import the pool with a "zpool import" but again everything seems broken and I can't start the services or the array properly.

Diags attached.

For the record this has all been with 7.2. This was a fresh build with 7.2 with a whole new pool and I was busy restoring data back. I had made some of it live so I'd really rather not lose data since the last backup (last night) and I really don't want to start the restoration seeds all over again (~30TB already restored before this happened)

ixios-diagnostics-20251118-1123.zip

UPDATE: Not fixed, but functional. Sort of.

Did a bit of faffing around. After starting the array and stopping it a couple of times it finally exported properly on array stop. So I then went in and started the array one more time and it seems to be up and running. All the Dockers are running and the shares appear to be working. However, the GUI of UNRAID still thinks the array is not running so I can't manage shares or anything. Going to "Shares" tab gives me "There are no mounted array or pool disks - cannot add shares."

Assistance still required to get myself out of this pickle... but at least my array is functioning again even if not working well.

Post new diags please.

Done

ixios-diagnostics-20251118-1156.zip

It looks like the pool was imported manually; the GUI won't recognize that.

Stop the array, reboot the server, and post new diags after array start, don't import the pool manually.

Done. Note that after just starting the array the pool is unusable and not imported.

ixios-diagnostics-20251118-1204.zip

And again, after stopping the array the pool is apparently mounted. Ran another diagnostic with array stopped.

ixios-diagnostics-20251118-1205.zip

You must have something importing the pool, a plugin or a script. Reboot in safe mode, and before starting the array, reimport the pool:

on main click on the first device for that pool and then "remove pool"

back on main, create a new pool with the same name and number of slots

assign the pool device(s), leave the filesystem set to auto

start the array to import the pool and post new diags.

I can't leave the filesystem set to "Auto". It doesn't give me the ability to add subpools unless I set it to ZFS and without the metadata the pool is dead anyway.

I did isolate a shutdown script that was importing the pool and have removed that... so that mystery's solved at least.

Edited November 18, 2025Nov 18 by Sinister Crayon

1 minute ago, Sinister Crayon said:

I can't leave the filesystem set to "Auto".

Yes, you can, and you must. Subpools devices are assigned together with the other devices; all together, they will be sorted at array start.

Done. Pool inaccessible. Did not import.

ixios-diagnostics-20251118-1243.zip

Pool is failing to import:

Nov 18 12:42:31 Ixios root: cannot import 'ixios' as 'ixios': I/O error

Nov 18 12:42:31 Ixios root: Destroy and re-create the pool from

Nov 18 12:42:31 Ixios root: a backup source.

Curious that just zpool import appears to be working, possibly it's using a different device. Post the output from:

zpool import -f -m -N -d /dev/sdm1 1247441232252185846 ixios

and if that fails try

zpool import -f -m -N -d /dev/sdn1 1247441232252185846 ixios

P.S. The previous zpool status showed that the pool has errors, so it may be related to that.

errors: 1 data errors, use '-v' for a list

root@Ixios:~# zpool import -f -m -N -d /dev/sdm1 1247441232252185846 ixios

cannot import 'ixios' as 'ixios': I/O error

Destroy and re-create the pool from

a backup source.

root@Ixios:~# zpool import -f -m -N -d /dev/sdn1 1247441232252185846 ixios

cannot import 'ixios' as 'ixios': I/O error

Destroy and re-create the pool from

a backup source.

root@Ixios:~#

And just zpool import ixios

BTW, I'm aware of the errors. That was a DB file that got corrupted due to the sudden power off. Non-critical file and replaceable. 

root@Ixios:~# zpool import ixios

root@Ixios:~# zpool status -v

pool: ixios

state: ONLINE

status: One or more devices has experienced an error resulting in data

corruption. Applications may be affected.

action: Restore the file in question if possible. Otherwise restore the

entire pool from backup.

see: https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-8A

scan: scrub canceled on Mon Nov 17 13:47:12 2025

config:

NAME STATE READ WRITE CKSUM

ixios ONLINE 0 0 0

raidz2-0 ONLINE 0 0 0

sdm1 ONLINE 0 0 0

sdn1 ONLINE 0 0 0

sdc1 ONLINE 0 0 0

sdj1 ONLINE 0 0 0

sdf1 ONLINE 0 0 0

sdk1 ONLINE 0 0 0

sdg1 ONLINE 0 0 0

sdh1 ONLINE 0 0 0

sde1 ONLINE 0 0 0

sdl1 ONLINE 0 0 0

sdd1 ONLINE 0 0 0

sdi1 ONLINE 0 0 0

special

mirror-1 ONLINE 0 0 0

nvme1n1p1 ONLINE 0 0 0

nvme3n1p1 ONLINE 0 0 0

logs

mirror-15 ONLINE 0 0 0

nvme0n1p1 ONLINE 0 0 0

nvme2n1p1 ONLINE 0 0 0

errors: Permanent errors have been detected in the following files:

/mnt/ixios/appdata/syncthing/index-v2/folder.0002-bchrqmal.db

Pool should also import using the devices. Try scrubbing it, and then try letting Unraid import it again. If it still fails, I recommend backing up and recreating the pool.

Forgot to mention, after a manul import, you should be able to access the pool data under /mnt/ixios

A scrub is probably going to take 24 hours or more and I think is unlikely to resolve the issue. The fact that it imports means that it's usable... I can literally read data from it. 

I tried booting up an Ubuntu live USB stick and I can import the pool there too. No issues. Are you telling me there's no effective way to just import an existing ZFS pool into unRAID? Imagine for a moment this is an existing ZFS pool (for example from TrueNAS) that I want to import into a new unRAID install; how would I do that? 

Is there any configuration on the USB stick I should be looking at? Some saved value that has gotten broken somewhere? 

Wait a minute, I missed before that you didn't assign the subpool devices. This was the command Unraid was using

Nov 18 12:42:31 Ixios emhttpd: shcmd (288): /usr/sbin/zpool import -f -m -N -d /dev/sdm1 -d /dev/sdn1 -d /dev/sdc1 -d /dev/sdj1 -d /dev/sdf1 -d /dev/sdk1 -d /dev/sdg1 -d /dev/sdh1 -d /dev/sde1 -d /dev/sdl1 -d /dev/sdd1 -d /dev/sdi1 1247441232252185846 ixios

You must assign all the devices, including the subpool devices, or it will fail, I mentioned that before.

You literally told me I "must" use "Auto" filesystem. If I select Auto filesystem then I can't assign subpool devices. 

Just now, Sinister Crayon said:

If I select Auto filesystem then I can't assign subpool devices. 

35 minutes ago, JorgeB said:

Subpools devices are assigned together with the other devices; all together, they will be sorted at array start.

Context here; your response to my earlier concern was

"Yes, you can, and you must. Subpools devices are assigned together with the other devices; all together, they will be sorted at array start"

I thought I was clear, but sorry if I wasn't, the pool must have 16 slots, you must assign all the devices, including the subpool devices, all together, and leave the filesystem set to auto, they will be automatically sorted at array start.

repeat the pool import.

Thank you. No that wasn't clear but indeed this seems to have resolved the problem. Going to try a couple of reboots to be sure before I mark it as a solution but it's looking good at the moment.