Help - Read Error during Disk Rebuild

[graywolf]

Diagnostics attached.

So is the rebuild messed up and I need to try again?

Replaced a 6TB drive with an 8TB drive, currently 3.40 TB into the rebuild

tower-diagnostics-20211201-0125.zip

[trurl]

Why did you decide to replace disk11? Disk20 has SMART warnings on the Dashboard. Do any of your other disks have SMART warnings on the Dashboard page? Do you still have the original disk11? Has anything been written to your server since disk11 was disabled or replaced?

 

Why are you running such an old version?

 

You really should have dual parity with that many data disks.

[graywolf]

Disk11 showed error on Dashboard with 48 reallocated sectors and a bunch of raw read errors. Since I wanted to replace a 6TB with an 8TB drive I picked that one.  Disk6 has 1 reallocated sector.  Disk20 showed as Healthy on the Dashboard before the disk11 rebuild.

 

I have the original disk11, but there have been writes to the array.

 

[trurl]

Are the read errors increasing?

[graywolf]

Disk rebuild just finished 18hrs 35mins 35secs 128 errors

Disk11 returned to normal operation

Current diagnostics attached

Disk20 showing 1 Current Pending Sector 

tower-diagnostics-20211201-1346.zip

[trurl]

Looks OK. You should also monitor attributes 1 and 200 on WD disks, click on a disk to get to its page.

[trurl]

A filesystem check on the rebuild might be a good idea, but its probably OK.

 

12 hours ago, trurl said:

You really should have dual parity with that many data disks.

 

[graywolf]

So even with the 128 Read errors on disk20 the rebuilt disk11 should be good?

What should I do about the 1 Current Pending Sector on disk20?

 

I'll have to get another 12TB drive for dual parity, I'll work getting another one

 

What would be best approach in future prior to replacing parity or data drives?

Ensure dockers don't stop

Ensure nothing writes to disks (downloads, etc)

What else?

 

Edited December 1, 2021 by graywolf

[trurl]

A pending sector should become a reallocated sector when it is rewritten, disks have extra sectors just for that purpose. The trick is getting it rewritten.

 

Maybe building parity2 will make it consider rewriting the pending sector on that disk. We can check SMART for that disk then.

[JorgeB]

16 hours ago, graywolf said:

So even with the 128 Read errors on disk20 the rebuilt disk11 should be good?

There's could be some data corruption, if those read errors coincided with data on the rebuilt disk.