Help : Parity check finding 88 errors. What to do?

[DivideBy0]

So I am in a similar situation with the OP.  My power went off 2 or 3 times yesterday and I learned the hard way that my UPS batteries are kaput, done. I am getting new batteries as we speak.  Because of that my UnRaid kept bouncing with dirty shutdowns and it came back with an automated parity check. It was about 40% done and I noticed 2 errors I think sync and is perhaps due to the unclean shutdowns.  U stooped the parity check and manually start another one with "Write corrections to parity" enabled.  That's the right procedure, correct?  Anything else I should do? I hope there won't be many errors, but who knows.  Otherwise the disks are healthy and no other damage that I can see in logs from these hard / unclean reboots.

[trurl]

After correcting parity check, run a non-correcting parity check to confirm there are zero sync errors. Don't reboot between the parity checks, because if there are remaining sync errors you will need to get diagnostics so we can compare the parity checks.

[DivideBy0]

34 minutes ago, trurl said:

After correcting parity check, run a non-correcting parity check to confirm there are zero sync errors. Don't reboot between the parity checks, because if there are remaining sync errors you will need to get diagnostics so we can compare the parity checks.

 

Well funny you say that. I think when my second power snag hit, the UnRaid was in middle of the automated parity check after the first power failure forced reboot.  How is that into the mix now?  Should I stop this correcting parity check and do a diagnostic first?  Or let this one finish, do a non-correcting one after and if I don't have any more errors then I am good to go?  I hate when this happens and you think your UPC will cover your butt

Edited January 18, 2023 by johnwhicker

[trurl]

What we really need is the syslog with the correcting parity check, and the syslog with the following non-correcting check. Reboot is OK as long as these logs are saved before rebooting.

[DivideBy0]

Ok I will save the syslogs for correcting and none-correcting. I will do the full diagnostic as well just in case. 

 

So far is 100% into correcting parity check and I only have 2 errors as per below. Saved the syslog and diagnostic, rebooted and started a non-correcting parity check.  Will keep you updated.

 

From #dmesg I get this

[104634.736525] md: recovery thread: P corrected, sector=4362239800

[119130.066796] md: recovery thread: P corrected, sector=6442670392

 

And from cat /var/log/syslog.1 I get this 

Jan 18 00:07:47 NAS-UNRAID-1 kernel: md: recovery thread: P corrected, sector=4362239800

Jan 18 04:09:23 NAS-UNRAID-1 kernel: md: recovery thread: P corrected, sector=6442670392

Edited January 19, 2023 by johnwhicker

[DivideBy0]

@trurl Well the non-correcting check was finished and no errors found.  That means I am good? The correcting check found and fixed 2 sync errors as per above post.  Anything I need to worry about it?  

 

I am in process of replacing my UPS batteries so I should be covered for the next power hick up.

[trurl]

👍