Cache pool disks Unmountable: Unsupported or no file system [SOLVED]

[Raf]

Hi,

Using version 6.12.3.

I got an a bunch of emails from my server saying: "fstrim: /etc/libvirt: FITRIM ioctl failed: Input/output error" every hour. Once I noticed this I logged in and it appeared that all my dockers were stopped and the docker image was 110GB but the size is only set to 80GB in settings. I disabled the docker service and increased the size to 200GB. Started the docker service but it said that docker is not started. I rebooted the server. When it came back on line now my cache pool disks where the docker is located say "unmountable: unsupported or no file system", they show up in the cache pool slots but are not mounted and I have an option to format them (which obviously I don't want to do). Can you please help, I greatly appreciate your help!

 

Edited August 23, 2023 by Raf
remove attachment

[itimpi]

Making the docker image file larger than the default of 20GB is rarely required.   If you keep getting it filling up it nearly always means you have a container writing to a location that is internal to the container which instead should be mapped to a location external to the container.

[JorgeB]

If the log tree is the only problem this may help, type:

btrfs rescue zero-log /dev/sdf1

Then re-start array.

 

[Raf]

That worked! Thank you @JorgeB!

 

Before I got your solution, I did mount the drive to a temp location based on one of your previous posts and copied the data to the array and was planning to just recreate the pool and copy the data back. Then I unmounted the temp folder. The command you gave above worked like a charm though. My docker image seemed to be corrupt though because the docker would not start. I just deleted it and recreated it. All good now. THANK YOU SO MUCH!

 

Just out of curiosity, what log from the diagnostic indicated that the problem might be the log tree?

[JorgeB]

Aug 22 16:08:11 Tower kernel: BTRFS: error (device sdf1) in btrfs_replay_log:2414: errno=-5 IO failure (Failed to recover log tree)

 

Note that this issue has been known to sometimes reoccur, if it does the same command should work again, but if that happens recommend reformatting the pool.