Confused about how to remove a disk, but also add / replace another

[Jacko_]

Not really sure what the best approach is for sorting this so thought i would ask for a bit of advice.

 

I had a drive (Disk 3) that looked like it was starting to fail (bad sectors), so i replaced it (Disk 6).

 

I've still got a shadow of Disk 3 in the array devices which i would like to remove but i'm a little apprehensive about doing so, in case i lose data - will it delete data from the parity / what is the best way to remove this stub without losing data?

 

Secondly i have another drive (Disk 5) which is now showing bad sectors, which i have bought another drive to replace.

I have just started a preclear on the new drive to replace Disk 5, so this will take a little while.  I was wondering if this was still the most up to date process before adding a drive to the array and what the best process is to replace a failing drive.  I guess i could just pull it, add the new drive into the Disk 5 slot and let the array rebuild, but this would put me at risk should parity drive fail during this process.  I think this is why i didn't chose this way to replace Disk 3 at the time.

 

I'm hoping to then use these two 12TB drives in another server once i have removed them from this one - the plan is to use them for a NVR project, if i get the time - maybe spin up another unraid server for this and have something i can play around with a little more without fear of breaking it / losing a lot of data.

 

Thanks

[trurl]

1 hour ago, Jacko_ said:

Not really sure what the best approach is for sorting this so thought i would ask for a bit of advice.

Really wish you had asked for advice before doing anything at all. Please don't do anything else without further advice.

 

1 hour ago, Jacko_ said:

I had a drive (Disk 3) that looked like it was starting to fail (bad sectors), so i replaced it (Disk 6).

That wasn't a replacement. That was adding a new disk6 and not doing anything at all about the disabled disk3. If you didn't have dual parity, it wouldn't have even let you add a disk when another disk was already disabled.

 

According to that screenshot, disabled/emulated disk3 is mounted and has 4.49TB of data. You have to rebuild disk3 or you will lose that 4.49TB of data on disk3.

 

  

1 hour ago, Jacko_ said:

this would put me at risk should parity drive fail during this process.  I think this is why i didn't chose this way to replace Disk 3 at the time.

Just digging a deeper hole, especially since another disk is now showing problems. Parity contains none of your data, and parity by itself can't recover anything.

 

https://wiki.unraid.net/Manual/Overview#Parity-Protected_Array

 

All disks are required to rebuild a failed disk. So disk5 failing is actually worse than parity failing, since disk5 has some of your data (unlike parity) and disk5 will be involved in rebuilding disk3 (or rebuilding parity if you actually want to remove disk3 from the array instead of rebuilding it.)

 

Fortunately, having dual parity will help you get out of this hole (that it allowed you to get into).

 

Not going to make any suggestions until you...

 

Attach diagnostics to your NEXT post in this thread.

[trurl]

@Jacko_

19 minutes ago, trurl said:

Please don't do anything else without further advice.

  

19 minutes ago, trurl said:

Attach diagnostics to your NEXT post in this thread.

 

[Jacko_]

Hi trurl,

 

Here's the diagnostics.  Ok understood I won't do anything for now.  Very glad of your assistance on this and I'll ready the wiki.

tower-diagnostics-20230101-0844.zip

[trurl]

Those diagnostics suggest your monthly parity check started at 3am and is still underway. Is that correct?

 

Is the disk originally assigned as disk3 still attached? Was it sdg 9RKB9HLL?

 

 

[Jacko_]

Yes that's right on both points.

[Jacko_]

Just to add, the new drive is at 4% Zeroing.

 

I decided not to stop parity check, thought best to see what you suggest as a remedy.

 

Thanks.

[trurl]

Original disk3 failed extended self-test so it should definitely be replaced.

Serial Number:    9RKB9HLL
# 1  Extended offline    Completed: read failure       90%     17004         1341797320

 

Disk5 recently passed extended self-test and those SMART attributes aren't terrible unless they begin to increase.

 

Do any other disks have SMART warnings on the Dashboard page?

 

If parity check is going OK probably disk5 is fine, since original disk3 is no longer in the array it isn't involved in parity check. Just let parity check continue.

 

Probably disk5 will be OK at least until you rebuild disk3 to that preclearing disk.

 

 

 

 

[Jacko_]

Parity check has completed ok, just waiting on this new disk to be ready.

 

[Jacko_]

PreClear has completed.

[trurl]

Post new diagnostics just to check if all is still working well before rebuilding disk3.

[Jacko_]

Here you go.

 

 

tower-diagnostics-20230103-1423.zip

[trurl]

Emulated disk3 still mounted, no obvious I/O errors in syslog. Should be good to rebuild disk3 to that new disk.

 

https://wiki.unraid.net/Manual/Storage_Management#Normal_replacement

 

Ask if you have any questions

[Jacko_]

So stop the array, and select the new disk to replace disk 3, is that right?

 

 

[trurl]

Stop array, assign replacement to the disk3 slot, start array to begin rebuild.

[Jacko_]

Good, because that's what I did. I did it in maintainance mode to be sure I don't run in to any problems.  Any way to check on the status of the rebuild?  Drive is currently showing as emulated.  Perhaps it will just change to active once complete.

[trurl]

In Main - Array Devices, lots of Writes to rebuilding disk, lots of Reads from all other disks, zeros in Errors column. Main - Array Operation should show rebuild progress.

[Jacko_]

I cam see reads from all but disk 3 and writes to only disk 3, just not sure how to check progress as I would need to take out of maintenance mode once completed.

 

Thanks for helping

[trurl]

1 hour ago, Jacko_ said:

not sure how to check progress as I would need to take out of maintenance mode once completed.

I've never done rebuild in Maintenance mode. If you do it in Normal mode, there is no reason you can't continue to use the array while rebuilding, it's just that reading or writing the array will impact rebuild performance and vice-versa. I sometimes use the array a little during rebuild or parity checks. Downloads to a cached share, for example, shouldn't impact anything if cache has space for the download and mover doesn't run during rebuild.

 

Do you mean you don't see anything in Main - Array Operation to indicate how much progress rebuild is making and how fast it's going?

[Jacko_]

Sorry, I can see it.  I was struggling on my phone.

 

It's only 2% completed so I'll report back when it's finished.

[Jacko_]

Array is showing an error, i think this might be to do with Disk5 - showing more issues than before. 

 

 

 

tower-smart-20230104-0846.zip tower-diagnostics-20230104-0848.zip

[JorgeB]

That is normal until disk3 finishes the rebuild.

[Jacko_]

Thanks JorgeB,

 

I thought that might be the case - nothing showing within fix common problems to indicate other issues but defiantly have more errors on Disk5 now.  Once the rebuild has completed (now at 82%) i'll do some tests on Disk5 to see if i should replace it.

 

 

 

 

[Jacko_]

Hi both.

 

The disk3 rebuild has completed successfully so that's great news.  I'm just doing an extended test on Disk5 and i'll share the results when completed.  I might get a disk bought in readiness to replace it, as i think it would be a good idea to pop these into another chassis.  I'll have a think.

 

Many thanks once again

 

 

 

 

[Jacko_]

Ok so here are the results of running an extended test on Disk5.  I think given the 187 Reported_Uncorrect 1 showing then it's due for replacement...?

 

tower-smart-20230106-0847.zip

Edited January 6, 2023 by Jacko_