This morning I have 2 errors from Fix Common Problems and I'm not sure what I need to do. Thought I would ask before I jump and assume.

Error 1:

Unable to write to cache - Drive mounted read-only or completely full. Begin Investigation Here: Unraid Main

Error 2:

Unable to write to Docker ImageDocker - Image either full or corrupted. Investigate Here: Docker Settings

 

Here is the disk log from the cache.

Mar 16 00:00:40 UnNAS kernel: BTRFS error (device sdd1): unable to find ref byte nr 128548089856 parent 0 root 5 owner 265457 offset 85291008
Mar 16 00:00:40 UnNAS kernel: BTRFS: error (device sdd1) in __btrfs_free_extent:6802: errno=-2 No such entry
Mar 16 00:00:40 UnNAS kernel: BTRFS info (device sdd1): forced readonly
Mar 16 00:00:40 UnNAS kernel: BTRFS: error (device sdd1) in btrfs_run_delayed_refs:2935: errno=-2 No such entry
Mar 16 00:00:40 UnNAS kernel: BTRFS error (device sdd1): pending csums is 66195456

I assume this is the cause of the unwriteable docker but I'm unsure of how to fix it. I'm still only a month green at this.

I've attached my log file as well.

Thanks.

unnas-diagnostics-20200316-0740.zip

Cache filesystem is corrupt, best to backup, re-format and restore the data.

4 minutes ago, johnnie.black said:

Cache filesystem is corrupt, best to backup, re-format and restore the data.

Is it safe to backup what is currently there since the filesystem is corrupt?

Yes, don't try to backup the docker image, it's easy to recreate, also see here if the fs goes unmountable before the backup is done.

3 minutes ago, johnnie.black said:

Yes, don't try to backup the docker image, it's easy to recreate, also see here if the fs goes unmountable before the backup is done.

Ok, thanks. I couldn't run krusader so I'm manually copying the cache drive to a local computer now. I also have an auto backup from yesterday as well. I'll wait for the backup to finish, hopefully before anything happens, and then follow the instructions you provided.

Well, the backup finished so I attempted to stop the array but it got stuck in a loop trying to unmount the disk shares. Gave it about 20 minutes then opened a terminal and ran "fuser -km /mnt/user" but that didn't help. I remembered I was sshed in to the cache so I ran "fuser -km /mnt/cache" but that didn't help. When to run it for disk1 but accidently hit enter after "fuser -km /mnt/" and my terminal connection dropped and the webgui is unavailable.

 

Guess I'll have to wait till I get home tonight and do a hard reboot.

BTW, figured I would ask. How do I figure out what may have caused the file corruption on the cache drive? I would like to avoid it from happening again if possible.

Difficult to say for sure, if it happens regularly it's likely a hardware problem, like bad RAM.

8 minutes ago, johnnie.black said:

Difficult to say for sure, if it happens regularly it's likely a hardware problem, like bad RAM.

Thank you

Daughter is home today so I had her reboot my server. Aside from the forced parity check, everything came back up normal, including docker, all it's containers and the cache doesn't show any errors.

 

Is there anything I should look at or maybe rebuild the cache anyway? Parity check will take about 14 hours. Should I just let things run their course and see if I get a file system corruption again?

20 minutes ago, lviperz said:

Is there anything I should look at or maybe rebuild the cache anyway?

I would recommend doing that, as it's likely to happen again soon.

Thanks again.