[Solved] Disk read errors while rebuilding

[gerhard911]

I was upgrading a 10TB disk to a 14TB disk (#6) and during the rebuild hundreds of thousands of read errors were reported on a 4TB disk (#13). The rebuild seemed to complete successfully but I can see that the rebuilt drive is missing most of the data it should have. Additionally, disk13 is now in a disabled state.

 

I have the original 10TB disk6 intact. No data has been written to the array since the rebuild so parity should reflect the state of the array before rebuilding.

 

So it seems to me that there should be some way to put the original disk6 back in and tell unRaid to rebuild disk13. Can this be accomplished or am I going to lose data?

 

TIA for any advice or assistance.

Edited September 7, 2020 by gerhard911
spelling

[JorgeB]

Please post the diags: Tools -> Diagnostics

 

[gerhard911]

unraid1-diagnostics-disk13 disabled-20200907-1014.zip

[gerhard911]

Here is the syslog during the rebuild.

unraid1-14TB rebuild disk errors-syslog-20200904-1817.zip

[JorgeB]

Disk13 dropped offline, looks more like a controller problem, SASLP are not recommended for a long time, but since there's no SMART reboot and post new diags, but note that the rebuilt disk will be mostly corrupt due to those errors.

[gerhard911]

After reboot disk13 still disabled. I did not restart the array.

 

I know my original disk6 has all data intact and parity should be in sync with that disk. Again, is there a way (new config ?) to rebuild disk13 from parity with my original disk6 ?

unraid1-diagnostics-reboot-20200907-1104.zip

[JorgeB]

33 minutes ago, gerhard911 said:

After reboot disk13 still disabled.

That's expected, once a disk is disabled it stays disable until rebuilt.

 

Disk13 looks fine, still good to check that it mounts correctly with UD in read-only mode, if yes do a new config with old disk6 and resync parity, then repeat the replacement.

 

You should also replace the SATA cable for disk14 since there are a few CRC errors on the previous diags.

 

Also good idea to replace that SASLP ASAP.

[gerhard911]

OK, I apologize I don't know what 'with UD in read-only mode' means. Is that starting the array in 'Maintenance Mode'? How do I confirm it has mounted correctly?

 

I'm also a bit confused about the details of the new config process. Please bear with me and confirm I have this right.

 

I replace the improperly rebuilt 14TB with the original 10TB. Then I do a New Config using what 'Preserve current assignments' ? It seems I would check both Parity slots and Data slots.

 

I follow the New Config by rebuilding data on disk6. Is this what you referred to as a parity resync?

 

Following that I can reattempt my 10TB to 14TB upgrade.

 

Thanks for your expertise!

[JorgeB]

16 minutes ago, gerhard911 said:

OK, I apologize I don't know what 'with UD in read-only mode' means. Is that starting the array in 'Maintenance Mode'? How do I confirm it has mounted correctly?

Sorry, I sometimes forget not everyone is aware of everything related to Unraid, UD is the Unassigned Devices plugin, you can install it from the apps tab, check the first couple of posts on the support thread, that and the included help should give you an idea of how to use it, if still doubts ask.

 

Note that you need to unassign the disable disk and start the array to make Unraid "forget it" before it can be used with UD.

 

21 minutes ago, gerhard911 said:

I replace the improperly rebuilt 14TB with the original 10TB. Then I do a New Config using what 'Preserve current assignments' ? It seems I would check both Parity slots and Data slots.

Yes, new config, preserve everything, re-assign original 10tb if needed, array should look the same as it was before the upgrade, then start it to re-sync parity.

 

22 minutes ago, gerhard911 said:

I follow the New Config by rebuilding data on disk6. Is this what you referred to as a parity resync?

No, new config is to rebuild parity based on the old disk6, assuming disk13 is also looking correct, to bring the array back to the same state it was before the upgrade, after that is done and parity is valid you repeat the disk upgrade.

[gerhard911]

OK, thanks for the followup. We are assuming that even though disk13 had a bunch of read errors and got disabled, that the data on it will be fine if it mounts with UD in read only mode.

[JorgeB]

Yes, the problem appears to have been caused by the controller, not the disk.