A few days ago I had a very dirty hard shutdown while Mover was running on my btrfs cache pool (2 nvme drives 1tb & 500gb). On reboot I noticed that the pool devices were unmountable. I ran btrfs check and found a corrupt block. After several attempts to repair it I was able to remount the drives and pull a backup copy of my appdata and VM ISOs with some unreadable files that i skipped over (mostly Plex posters).

I deleted my docker image and reinstalled my apps and copied appdata back. Everything seemed fine at first until I started geting file sytem is read-only errors and very slow unmounting and mounting disk times.

I decided to remove, clear and reformat my cache pool devices and copy my backed up appdata back over. Again everything was up and running for a few minutes until the read-only error came back. I pulled diagnostics right before another hard shutdown and am currently running memtest with 0 errors.

Is there anything I can do with the filesystem besides replacing the RAM and NVME drives?

laniakea-diagnostics-20251111-2229.zip

Unlikely that the NVMe device is the problem, I see that your RAM is overclocked. Start by correcting that, since Ryzen with overclocked RAM is known to corrupt data:

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/page/2/#findComment-819173

Thanks, JorgeB. I was running the D.O.C.P. (edited from XMP) profile out of habit. Server is my old gaming platform. I will set it back to 3200. This has been running solid for years with no issues until this week with the dirty shutdown during Mover job. Is there anything else in the diags?

Where can I go in the diagnostics to view my system info?

Edited November 12, 2025Nov 12 by CaptainTrap

Everything has been running fine since my last post (two days ago) PMS and other dockers working as expected. However early this morning Docker service failed to start and I saw BTRFS corruption errors again! Diags attached. Any help is appreciated.

laniakea-diagnostics-20251114-0819.zip

The only issue I'm seeing for now is the pool going read only dur to running out of space

Nov 14 04:02:20 Laniakea kernel: BTRFS: error (device nvme0n1p1 state A) in remove_from_free_space_tree:865: errno=-28 No space left

Nov 14 04:02:20 Laniakea kernel: BTRFS info (device nvme0n1p1 state EA): forced readonly

Note that while btrfs can for some time use the full capacity of a RAID0 pool using two different-sized devices, metadata is set to RAID1 by default, so it needs some space for that on the other device, which you currently don't have.

You can use single profile instead of RAID0, since that will redistribute the writes differently, always going to the device with the most free space first, or change the metadata to single (no redundancy . I would recommend the former, but note that in either case you will need to free up some space before attempting to convert the profile, or it will also fail due to not enough space.

Diasble docker and VM services, restart the array, the pool will hopefully mount read/write, move some data to free up space, convert the filesystem to the desired profile.

I disabled docker and VM and stopped array but its stuck unmounting:

/mnt/cache: target is busy

/mnt/disk5: target is busy

You will likely need to force a reboot.

Just wanted the thank you for your support @JorgeB in holding my hand through this process. Everything seems to be running fine after reading your other posts on similar BTRFS corruption topics and your suggested resolutions as well as responding to my own queries. Thanks!