As the title says, my cache-vm pool which consists of 2 SATA SSD Disks running at BTRFS RAID1

• /dev/sdg

• /dev/sdf

/sdf is a bad disk I wished to remove. I did the following:

1. in cache-vm pool settings, I changed the balance setting to "Convert to Single mode"

2. Waited for few minutes. refreshed the pool settings page, nothing changed (pool still at RAID1)

3. restarted the array

4. pool now shows that it's unmountable

My goal is to remove /sdf and keep /sdg for a now while waiting for my new set of disks to arrive.

Any ideas on how to resolve the unmountable issue and re-attempt converting the pool from RAID1 to single mode?

Please let me know if I'm missing more information :)

Diagnostics attached

ishtarcommander-diagnostics-20250723-0322.zip

I noticed that /sdf was attached in unsassigned disks so I

1. Unmounted the disk

2. Stopped the array

3. reseated both /sdg and /sdf

4. Started the array

Balance is currently running. I'll let it finish, and then try converting RAID1 again to Single Mode

Post new diags after the balance finishes.

7 hours ago, Ishtar said:

I'll let it finish, and then try converting RAID1 again to Single Mode

Note that you will only be able to balance to the Single profile after you have removed the second drive. You cannot remove ANY drive from a multi-drive pool in Single profile without breaking the file system.

2 hours ago, itimpi said:

You cannot remove ANY drive from a multi-drive pool in Single profile without breaking the file system.

Yes, good point.

On 7/23/2025 at 6:35 PM, itimpi said:

Note that you will only be able to balance to the Single profile after you have removed the second drive. You cannot remove ANY drive from a multi-drive pool in Single profile without breaking the file system.

As you guys mentioned, there's no way to remove a disk in Single Mode (and didn't understand what that mode is)

So I had another workaround instead and used this opportunity to move away from BTRFS due to accumulated bad experience over the years

1. stop docker service

2. shut down VMs and stop the service

3. stop the array/pools

4. open terminal and enter btrfs rescue zero-log /dev/sdf1

5. start the array

6. move the vmdata pool files to another pool (I chose appdata)

7. retarget folder shares to appdata

8. start the VMs and confirm it's working

9. shutdown the VMs, then stop the VM service, stop the array

10. wipe the vmdata pool, reseat the drives, and then rebuild it as ZFS mirror

11. start the array, move the files from appdata to vmdata, test VMs, enable docker

On 7/23/2025 at 4:59 PM, JorgeB said:

Post new diags after the balance finishes.

I attached new diags as of July 26 (despite being "resolved")

izanagi-diagnostics-20250726-2117.zip

Just one last question before closing this thread, what should be the proper method of removing a disk in a mirrored BTRFS pool?

Edited July 26, 2025Jul 26 by Ishtar
formatting

21 minutes ago, Ishtar said:

As you guys mentioned, there's no way to remove a disk in Single Mode

You could also have converted to raid1 profile, then removed a device, and the pool would automatically convert the remaining device to single.

15 hours ago, JorgeB said:

You could also have converted to raid1 profile, then removed a device, and the pool would automatically convert the remaining device to single.

I also found an old thread which still has the relevant information about BTRFS and other FS

FAQ for unRAID v6 - General Support - Unraid

closing this thread! :)