Hello all,

 

I've tried looking through the forums here for similar situations but I have been unable to find any.

 

I have a data disk that failed that needs to be replaced and I have already swapped out the failed drive with its replacement. The trouble is that at the same time, the 4 Bay USB enclosure I had these disks in also failed (I know, USB enclosures aren't recommended) and I decided to move all of the drives to my server box instead of having them external.

The trouble I now face is twofold.

 

1. Following the WIKI steps here: https://wiki.unraid.net/Replacing_a_Data_Drive, at step 2 I accidentally unassigned the wrong disk from the disk configuration. I should have unassigned disk 3, but instead have unassigned disk 1 and have rebooted.
   1. 
2. Since I have now moved the disks from the external enclosure to the server itself - connected via a PCI SAS card / mini SAS to SATA cables - the drives now show up as something different due to what I think is the "Display world-wide-name in device ID" setting and the move from enclosure to server. As you can see before, all of the disks were appended with the same number while in the enclosure and are now appended with their SN as expected.
   1. 

 

So to summarize - 1. Disk and enclosure failed. 2. unassigned wrong disk, 3. disks now show up under different IDs, 4. The failed drive - Disk 3, a 4TB Seagate - has been replaced with an 8TB Seagate.

 

I thought about just making a new config, but in this case I can't do that because I have a drive to rebuild. Does anybody have any suggestions for what I can do in this situation? Any and all help would be greatly appreciated.

Edited December 6, 20214 yr by Gevlleg
adding more info to summary, updating title to solved

1. New Config, assign all disks as they should be. You can assign the old or new disk to the slot to be replaced.
2. Before starting the array, check the box saying parity is valid AND check the box for Maintenance mode. This makes all disks accepted as a new array, but doesn't change existing parity or mount the disks, so nothing is written during mount that would need parity to change.
3. Stop the array
4. Unassign the disk you want to rebuild
5. Start the array with that disk unassigned. That makes the disk disabled.
6. Assign new disk and start the array to begin rebuild.

Ask if you have questions.

Thank you for the quick reply and succinct answer.

Just to be sure I understand completely, I want to clarify how this would be going in step 1.

 

The old disk that failed is no longer detected - clicks an insane amount when powered on and then I get nothing, so I can't assign that one anymore.

 

So what I would do for step one is assign all of the disks in the slots as follows:

 

1. 8TB (Original, but unassigned on accident)
2. 8TB (Original)
3. 8TB (Replacement for the failed 4TB)
4. 4TB (Original)

 

And from there, I continue down your list of operations. Do I understand you fully? Thanks again so much.

16 hours ago, Gevlleg said:

Do I understand you fully?

yes

At step 5 the unassigned/disabled disk should be emulated and its contents accessible as if it were there, unless the emulated disk is unmountable. If that happens we can repair the disk before rebuilding or after. Probably moot whether it is before or after since the original is unusable anyway.

 

Whatever you do, don't format anything. Format is NEVER part of rebuild.

Thanks trurl, I have started the rebuild process and am at 1% right now of about 717GB to rebuild. I appreciate your help and I will keep this post updated with the results!

The rebuild has completed successfully and all data seems to be intact. Thank you so much trurl, you were a wonderful help.