First time poster in the unraid forum, so any pointers would be appreciated!

A few weeks ago I ran into an error where my docker service was not starting. I did some digging online, and came across a page suggesting me to increase my vDisk size. I did that, and my server was normal for a few days. I ran into another error, this time (to my best recollection) was error 503 trying to start a docker container. I tried to restart my server, but the cache disk failed to unmount. I went ahead and rebooted the server and all was back to normal again.

Fast forward to today, when I noticed that I couldn't restart a docker container. I tried restarting my array instead, but saw that my cache drive was failing to mount with the error Unmountable: Unsupported or no file system. Reading the docs led me to run `btrfs check --readonly /dev/nvme0n1p1 which produced the following output. Most of these included the file names,

Output

Wrong key of child node/leaf, wanted: (255176, 108, 5472256), have: (255176, 108, 5488640)

root 5 inode 255127 errors 1000, some csum missing

root 5 inode 255176 errors 400, nbytes wrong

root 5 inode 255177 errors 2001, no inode item, link count wrong

unresolved ref dir 254514 index 630 namelen 5 name 17615 filetype 1 errors 4, no inode ref

root 5 inode 255178 errors 2001, no inode item, link count wrong

unresolved ref dir 254514 index 631 namelen 5 name 17621 filetype 1 errors 4, no inode ref

root 5 inode 255179 errors 2001, no inode item, link count wrong

unresolved ref dir 254514 index 632 namelen 5 name 17624 filetype 1 errors 4, no inode ref

root 5 inode 255180 errors 2001, no inode item, link count wrong

unresolved ref dir 254514 index 633 namelen 5 name 17626 filetype 1 errors 4, no inode ref

root 5 inode 255181 errors 2001, no inode item, link count wrong

... (there are quite a lot of these for different files)

The manual advised against running with the --repair flag as it could further damage the filesystem. I'm not very familiar with these errors so it would be helpful to know what steps to go after here. I'm mainly more hopeful to get my docker images back and not need to go through setting everything up again. Thanks for any pointers!

Attach Diagnostics ZIP to your NEXT post in this thread.

Diagnostics attached here:

800g6-diagnostics-20250604-2326.zip

Type btrfs rescue zero-log /dev/nvme0n1p1 then restart the array and post new diags.

Oh hey, that brought the drive back. Thanks so much! Would you mind explaining a little bit about what the errors/fix were?

Diags attached: 800g6-diagnostics-20250605-0644.zip

Looks OK for now, diags were showing that the log tree was the problem, and the command above zeroes it.

Note that this issue can sometimes reoccur, if it does, the same command should work again, but I would then recommend backing up and reformatting the pool, since there may be other issues.

Thanks for the explainer and the help.

I watched a few movies on my server today, but noticed some weird behaviour again (Jellyfin not saving what is watched, not loading images). I poked through the logs and saw a sea of BTRFS errors again.

I'm starting to wonder if there is a different root cause here... I checked the SMART status of the SSD and saw everything is normal. If a reformat is needed, are there steps I can take to save my docker images and containers? I know they don't actually store the data for my backups, but want to triple check to make sure.

800g6-diagnostics-20250605-2252.zip

Jun 5 06:45:21 800g6 kernel: BTRFS: error (device nvme0n1p1: state A) in btrfs_finish_ordered_io:3319: errno=-117 Filesystem corrupted

Jun 5 06:45:21 800g6 kernel: BTRFS info (device nvme0n1p1: state EA): forced readonly

Pool went read-only, recommend backing up and reformatinmg to avoid more issues.