Parity checks now showing errors

[shaunmccloud]

I've had unraid running for a couple of years now, never had a single issue when running a parity check. The latest parity check showed 900+ errors.  Ran a correcting parity check then another non-correcting one and the same amount of errors even though the correcting one fixed said errors.   Do I have a parity drive dying?

Diagnostics is attached.

bb-8-diagnostics-20230606-0905.zip

[JorgeB]

12 hours ago, shaunmccloud said:

another non-correcting one and the same amount of errors

It may be the same amount but they are in different sectors, also a lot of them sequentially to be a transient RAM issue, most likely culprits would be a controller or a disk, if it's a disk it can be a pain to find it, because you'd basically need to re-test after removing/replacing one disk at a time, if you have a different controller you could use I would start there.

[shaunmccloud]

7 hours ago, JorgeB said:

It may be the same amount but they are in different sectors, also a lot of them sequentially to be a transient RAM issue, most likely culprits would be a controller or a disk, if it's a disk it can be a pain to find it, because you'd basically need to re-test after removing/replacing one disk at a time, if you have a different controller you could use I would start there.

I can replace one of my controllers, but not both of them.  I am out of PCIe ports to swap out two controllers

[JorgeB]

Try just one, if the same swap with the other, note that you must run at least two checks to confirm if the issue is still there or not, 1st might still find errors, 2nd one can't.

[shaunmccloud]

19 minutes ago, JorgeB said:

Try just one, if the same swap with the other, note that you must run at least two checks to confirm if the issue is still there or not, 1st might still find errors, 2nd one can't.

Should have said, one of my SAS controllers is onboard.  So to be able to replace both of them I would have to get a 16 port controller.  But I will replace the one I can with my cross flashed H330 (IT mode) as a start.  If I have to, I will pick up a new 16 port HBA and new cables (since most 16 port I can find on eBay use the newer connector type).

Edited June 7, 2023 by shaunmccloud

[shaunmccloud]

@JorgeBSince I am not good at reading the logs, which drives had the issues?  I have 4 on one controller and 8 on another one right now.  Knowing which drives would help me narrow down which controller.

[JorgeB]

9 hours ago, shaunmccloud said:

Since I am not good at reading the logs, which drives had the issues? 

That's unfortunately not possible to know, why I mentioned that if it's a disk basically it will be a pain, because you'd need to remove one disk at a time and re-test until you find the culprit, like this guy had to do a while back:

 

 

[shaunmccloud]

 Swapped my addon HBA and ran a correcting parity check.  The email I got showed 490 errors, but the UI showed 0.  And looking at the log I do not see any issues like I did before.  New diagnostics attached.

bb-8-diagnostics-20230612-0702.zip

[JorgeB]

11 minutes ago, shaunmccloud said:

The email I got showed 490 errors

There's a known issue with the notifications, thought not clear what causes it since it's not reproducible, the log does show 0 errors, if the next one is also 0 I would consider yhe issue solved.

[shaunmccloud]

Ok, I do have a non-correcting running right now.  With the old controller, the parity check error count matched between the email & the UI.

[JorgeB]

3 minutes ago, shaunmccloud said:

With the old controller, the parity check error count matched between the email & the UI.

Pretty sure it's not the controller, that issue happens sometimes only.

[shaunmccloud]

1 minute ago, JorgeB said:

Pretty sure it's not the controller, that issue happens sometimes only.

Meant I think the old controller was bad since the two counts matched.  But with the new controller it does not match.

[JorgeB]

Agreed, if both checks complete without errors it could have been a bad controller.

[shaunmccloud]

And it might actually be one of my 4TB disks.

Event: Unraid Disk 8 error
Subject: Alert [BB-8] - Disk 8 in error state (disk dsbl)
Description: ST4000VN000-1H4168_Z301MZHT (sdh)
Importance: alert

 

[JorgeB]

That should not have caused sync errors, especially before it got disabled, but it's a possibility.

[shaunmccloud]

4 hours ago, JorgeB said:

That should not have caused sync errors, especially before it got disabled, but it's a possibility.

It did rebuild fine, running another parity check.  I should consider replacing my 4TB drives "soon" anyway.