Hi all.

I have 11 HDD's. I want to shrink my array to 8 HDD's.

I am attempting to remove 3 x 4 TB HDD from the array using the procedure as set out in the Unraid docs here: https://docs.unraid.net/legacy/FAQ/shrink-array/

All goes well until I restart the array, this is where my 2 x NVME cache disks and 1 x SSD pool disk get marked as unmountable unrecognized file system - the file system being also being marked as 'auto' when they were originally BTRFS.

I can cure the SSD issue by setting back it to BTRFS and it comes back fine.

The cache disks (2 x 1TB NVMe's) however, to do not come back online when I set them back to BTRFS + raid 1? They still are marked unmountable with unrecognized file system.

I have reinstalled a backup of my flash drive, so I am back up and running with the original system, but would like to know if there is anyway I can shrink the array by overcoming the cache pool disk issue as described above? Im sure I have shrunk my array in the distant past and not had the cache pool disk issue.

Any thoughts or help would be appreciated.

Edited June 1, 20251 yr by Vatoe
Typos

Please post the diagnostics after array start, also the output from btrfs fi show

Thanks for the reply.

This is the output from btrfs fi show:

Label: none uuid: dfa0320b-54f7-4176-9467-964c81fc02ad

Total devices 2 FS bytes used 219.47GiB

devid 1 size 931.51GiB used 247.03GiB path /dev/nvme0n1p1

devid 2 size 931.51GiB used 247.03GiB path /dev/nvme1n1p1

Label: none uuid: 1646dab7-c65f-4eb4-bfc5-043899e19337

Total devices 1 FS bytes used 118.48GiB

devid 1 size 931.51GiB used 124.02GiB path /dev/sdh1

Label: none uuid: d7706b04-ab94-4c77-b6cb-74870f06a749

Total devices 1 FS bytes used 37.60GiB

devid 1 size 80.00GiB used 66.07GiB path /dev/loop2

Label: none uuid: 064a4352-64bc-4aca-b600-41264734d038

Total devices 1 FS bytes used 1.08MiB

devid 1 size 1.00GiB used 126.38MiB path /dev/loop3

With the diagnostics, I'm assuming you need to see this when the error occurs? That is, I need to attempt the shrink array procedure again? If so, its evening where I am and the server is being used (streaming) still have to do it tomorrow morning my time.

Edited June 1, 20251 yr by Vatoe

11 minutes ago, Vatoe said:

I'm assuming you need to see this when the error occurs?

When the pool shows unmountable.

Just now, JorgeB said:

When the pool shows unmountable.

Yes okay thanks, I will do it in the morning or later tonight if my streamers get off before I sleep.

Hi Jorge

Managed to get the diagnostics done before I got to sleep.

See attached.

thanks.

edit:

added another btrfs output in the 'problem' state:

Label: none uuid: 30cbac5e-f6ea-4a4e-b4ee-56052fe17fe5

Total devices 1 FS bytes used 376.00KiB

devid 1 size 80.00GiB used 2.02GiB path /dev/loop2

Label: none uuid: 589c7630-502c-46d4-a101-7387c3f0a5ed

Total devices 1 FS bytes used 416.00KiB

devid 1 size 1.00GiB used 126.38MiB path /dev/loop3

Label: none uuid: dfa0320b-54f7-4176-9467-964c81fc02ad

Total devices 2 FS bytes used 219.52GiB

devid 1 size 931.51GiB used 248.03GiB path /dev/nvme0n1p1

devid 2 size 931.51GiB used 248.03GiB path /dev/nvme1n1p1

Label: none uuid: 1646dab7-c65f-4eb4-bfc5-043899e19337

Total devices 1 FS bytes used 118.48GiB

devid 1 size 931.51GiB used 126.02GiB path /dev/sdh1

darkmatter-diagnostics-20250601-2252.zip

Edited June 1, 20251 yr by Vatoe
added info

There's a strange emhttp error, please reboot first, try to reimport the pools and post new diags after array start:

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

25 minutes ago, JorgeB said:

There's a strange emhttp error, please reboot first, try to reimport the pools and post new diags after array start:

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

so I get this straight, im not doing the new config option?

I just did strictly per your instructions, therefore, I didn't use the new config option. Just removed the pool, re added it using 'auot', and the restarted the array. the cache came back up without issue. Attached are the diagnostics.darkmatter-diagnostics-20250602-1811.zip

Looks good, try now doing a new config, it should also work, but post new diags if it doesn't.

Okay that took a minute for me to reply, as while I was waiting for your last reply, I did try the new config option and the issue did re occurred with cache pool - unmountable. On this occasion however, the SSD was fine, being mountable. I then re-installed a backup copy of the flash drive (forgetting to get the diagnostics) and re did all the processes remove pool etc so as to get the a fresh copy of the diag's when the issue occurred again. However, this time I stopped the docker service before I restarted the array (essentially to save time if I was constantly starting a stopping the array) before restarting the array and to my supprise the cache pool came back online! I then stopped and restarted the array, restarting the the docker service in between and once again the cache pool came back online.

Weirdly never had an issue with the SSD being unmountable, like I did yesterday, it was fine all occasions.

So if this happens to anyone else and is reading this in the future, you may want to try stopping the docker service prior to restarting the array, get the cache pool online first, stop the array, restart the array and should be up and running. I don't think it was a co-incidence as I stated I did the have the unmountable issue on the previous try with the docker service being stopped.

Jorge, thanks for all your help!

Still not sure what was the problem, but glad to hear it's resolved.