I had a disc go unmountable after an unclear power cycle. I attempted running a filesystem check via the GUI; and both using -n or not resulted in a "FATAL ERROR: ALLOWED MEMORY SIZE OF XXXXXX BYTES EXHAUSTED". 

 

Running one via console; xfs_repair -nv completed so I ran xfs_repair -v and things stalled at process newly discovered inodes...
 

I quit the process and retried it. This time I let it go for a long time in case it was just taking a while, still stalling in the same spot.

 

So when that failed, I decided to try pulling the drive and replacing it with a new precleared disc to rebuild the array disc. Of course, it seemed to work until reboot when again this new disc is unmountable. I realize I skipped a few steps described here, but wanted to ask for help before I tried it since I did see some other posts stating rebuilding an unmountable disc will just end with another unmountable disc....

Attach diagnostics zip file to your next post in this thread.

Did you run the check from the GUI or the console ?

If you did it on the console, what exact command did you use ?

When I ran it from the GUI I got the out of memory fatal error code. 

 

When I ran 

xfs_repair -nv /dev/md5

in the console, it finished.

 

When I ran

xfs_repair -v /dev/md5

in the console, it hangs at  process newly discovered inodes...

 

 

dunraid-diagnostics-20221117-0822.zip

I assume you mean md3? There's no disk5, try rebooting and run it again from the console, if it still hangs it's likely an xfs_repair problem.

21 minutes ago, JorgeB said:

I assume you mean md3? There's no disk5, try rebooting and run it again from the console, if it still hangs it's likely an xfs_repair problem.

 

That is correct, thanks for catching my error.

 

I've rebooted and retried, same issue. 

 

Should I try to replace/rebuild the disc using the instructions here?

Rebuilding the disk won't help with a filesystem issue, you'd need to wait for a newer xfs_repair release or ask for help in the xfs mailing list.

1 hour ago, JorgeB said:

Rebuilding the disk won't help with a filesystem issue, you'd need to wait for a newer xfs_repair release or ask for help in the xfs mailing list.

 

Aside from that, am I looking at completely rebuilding the array?

Not the array, but one option would be to format that disk only, of course you would lose any data there, but if you have backups you could then just restore that data, another option, if there are no backups, would be using a file recovery util, like UFS explorer.

2 hours ago, JorgeB said:

Not the array, but one option would be to format that disk only, of course you would lose any data there, but if you have backups you could then just restore that data, another option, if there are no backups, would be using a file recovery util, like UFS explorer.

 

I'm fine with doing that, if I get no where with the xfs_repair mailing list. Is there a guide on doing that?

You basically just send an email describing the problem with the log, before doing it I would recommend running a file system check directly on the disk to rule out any md driver issues, unlikely it will make a difference but better to be sure

 

xfs_repair -v /dev/sdX1

Replace X with the correct disk letter and note the 1 in the end.

On 11/18/2022 at 3:24 AM, JorgeB said:

You basically just send an email describing the problem with the log, before doing it I would recommend running a file system check directly on the disk to rule out any md driver issues, unlikely it will make a difference but better to be sure

 

xfs_repair -v /dev/sdX1

Replace X with the correct disk letter and note the 1 in the end.

 

I'm sorry, I meant a guide on formatting that disc only.

 

I did run a file system check directly on the disc with the same results.

Start the array in normal mode and the format button will be next to the stop array button.