Parity Check Errors

[clowncracker]

I just started a parity check and there are already 261 errors found.  I do not have the parity check set to automatically fix errors.  Attached are my diagnostic files.  I ready that the Dynamix File Integrity plugin might help find the issue, but I thought the logs might provide information while the parity check is still running.

 

Edited July 26, 2023 by clowncracker
Removed Attachment

[itimpi]

2 hours ago, clowncracker said:

I ready that the Dynamix File Integrity plugin might help find the issue,

This is unlikely to find the issue.

 

Have you changed any drives recently?

[clowncracker]

1 hour ago, itimpi said:

This is unlikely to find the issue.

 

Have you changed any drives recently?

I swapped out both parity drives at the beginning of the month, but I ran parity checks then and didn't run into any issues.  I did have a power outage last night, but the shutdown should have been clean (server is attached to a UPS).

[clowncracker]

16 minutes ago, clowncracker said:

I swapped out both parity drives at the beginning of the month, but I ran parity checks then and didn't run into any issues.  I did have a power outage early this morning, but the shutdown should have been clean (server is attached to a UPS).

I'm looking at the logs and I don't see any mention of a shutdown.  Is there a reason why the server wouldn't have shutdown cleanly?  Here is the relevant log snippet:
 

Jul 25 06:12:45 clowncracker-NAS apcupsd[25426]: Power failure.
Jul 25 06:12:48 clowncracker-NAS apcupsd[25426]: Power is back. UPS running on mains.
Jul 25 06:12:49 clowncracker-NAS apcupsd[25426]: Power failure.
Jul 25 06:12:55 clowncracker-NAS apcupsd[25426]: Running on UPS batteries.

The next log files are at 7:42:37, so I don't think it shut down cleanly.  Here are my UPS settings, any reason why it didn't shut down cleanly?

 

Edited July 26, 2023 by clowncracker

[clowncracker]

UPDATE:

UDMA CRC error count on the Parity drive has gone from 0 to 3.  Not sure if that indicates if I should or should not write corrections.

[clowncracker]

18 hours ago, itimpi said:

This is unlikely to find the issue.

 

Have you changed any drives recently?

Parity check finished with 261 parity errors and UDMA CRC error count on the parity drive went from 0 to 3.  Should I just write corrections or is there something I can do to check the integrity of the files?

[itimpi]

If you have checksums then you can see if they match.   You can also compare against backups if you have them.   Other than that not much to do except correct errors.

 

CRC errors as such are only a consideration if they continue to increase.   They are typically connection issues relating to SATA/Power cabling and trigger retries so only really matter if they subsequently result in a read or write error.

[clowncracker]

23 minutes ago, itimpi said:

If you have checksums then you can see if they match.   You can also compare against backups if you have them.   Other than that not much to do except correct errors.

 

CRC errors as such are only a consideration if they continue to increase.   They are typically connection issues relating to SATA/Power cabling and trigger retries so only really matter if they subsequently result in a read or write error.

How would I go about checksums?  Even if I don't have them now, how can I implement them moving forward?  I wish I knew if the errors were coming from the parity drives or the data drives.

[itimpi]

6 minutes ago, clowncracker said:

How would I go about checksums?  Even if I don't have them now, how can I implement them moving forward?

 

If you use BTRFS (or ZFS) for the format of array drives then this is built into that file system type.    If using XFS then the Dynamix File Integrity can be used (although occasionally that seems prone to give false positives).