BTRFS volume fails every reboot, and sometimes randomly.

[witchy]

Hi folks,

My BTRFS experience hasn't been good up to now. I originally started with all BTRFS storage on TOS 5, and one by one they've failed until I have one volume left, and the only reason I've not binned it and remade it as ext4 is because it's 4TB in size. I'm not on TOS 6 after reading about the btrfs improvements.

More often than not, this volume fails to come back online after a restart, but also it's randomly disappeared when I've been looking at the TOS GUI. It's a file dump shared out with NFS and SMB, but it's not referenced that often.

SSH onto the TNAS and I can see that /Volume3 has been remounted as /var/subvols/8vEbTxkKvwc and is unreadable. So, umount that and run btrfsck against its logical volume which checks the filesystem and clears an error. Then I can remount it as /Volume3 and it's back online. For a while.

Anyone else get this issue?

Cheers
Witchy

[HeliaWu]

When you encounter this problem again, you can make an official appointment with our engineers to troubleshoot the problem remotely for you.

[witchy]

Having reread my post, I spotted a typo - it should say 'I'm NOW on TOS 6 after reading about btrfs improvements'

[TMroy]

For your data safety, you need to backup your data now. Because your file system can be failed again and not more able to be repaired. please check if your all drives are healthy. And avoid to forced power off or restart, or avoid any power outage.

[witchy]

Funny you should say that, this morning the web GUI wasn't reachable but SSH and backups were still working. Restarting nginx didn't help so I rebooted and guess what? /Volume3 has disappeared again. Thankfully I DID back this volume up to an external 4TB USB disk yesterday so I'm highly inclined to erase it and set it up as ext4 like the others which remain healthy. I can't see any error in dmesg or the logs in /var/log. There must be TOS errors logged somewhere.

I should say that the data is still there under /var/subvols/8vEbTxkKvwx/@.

For reference this is an F2-423 with 2x 8TB disks and a 256GB SSD as boot drive.

Storage Pool 1
- /Volume1 is the SSD (ext4)

Storage Pool 2
- /Volume2 is a Time Machine share (ext4), smb share only
- /Volume3 is the file dump (btrfs), smb/nfs share.

*edit* rebooting hasn't brought the web GUI back online. TNAS PC can't find it on the network even though I'm logged in over SSH using the same IP as it's always had. OK this is weird. It's a shame we don't have historical logfiles. I should also say that since yesterday the drives have been regularly accessed every few seconds even when there's no activity on the NAS. I've seen another post here saying the same thing.

Cheers
Witchy

[witchy]

Huh. It's only Edge that's decided the GUI's unavailable. Safari is connecting fine. I mean, it's not like Microsoft software to be buggy on MacOS (is there a sarcasm emoji?) but yesterday it was fine. Still got the btrfs issue though.