For a while I've had warnings from my parity disk (pending sector, reported uncorrect, reallocated sector). Was about to replace it since it is kind of old anyway but was in no rush to do so.

 

Yesterday one of my data disks went into disabled mode - contents emulated. Not sure why but I did check the hardware and replaced the sata cable. After a reboot i tried to run a parity sync/data rebuild since the data disk was now unrecognized. This is where it went bad.

 

The parity sync got stuck at 3% and i got flooded with warnings about the parity disk (offline uncorrectable, raw read error rate). Also had some light mechanical sounds thats not supposed to be there. Not sure which disk but my guess is the parity drive.  After a cancel and reboot the parity disk is now nowhere to be found in Unraid.

 

I have a replacement disk for the parity drive, how would I best go about trying to salvage this? Is disk 4 totally lost? What about disk 1-3?

2 minutes ago, Squamsh said:

For a while I've had warnings from my parity disk (pending sector, reported uncorrect, reallocated sector). Was about to replace it since it is kind of old anyway but was in no rush to do so.

Should have thought more carefully about that.

 

Attach Diagnostics to your NEXT post in this thread.

40 minutes ago, trurl said:

Should have thought more carefully about that.

 

Attach Diagnostics to your NEXT post in this thread.

Yeah. Lesson learned.

Diagnostics here

squamsh-unraid-diagnostics-20240829-1842.zip

50 minutes ago, Squamsh said:

Not sure why but I did check the hardware and replaced the sata cable.

You did this with the server running or powered down?

The question is whether the  data drive has really failed or just been disabled due to some other factor causing a write failure. .  It's SMART info looks OK.

13 minutes ago, Frank1940 said:

You did this with the server running or powered down?

I might have been a little bit reckless but I'm not THAT dumb :). Everything was powered off of course. It was hard shutdowns though since a soft shutdown didn't work.

 

14 minutes ago, itimpi said:

The question is whether the  data drive has really failed or just been disabled due to some other factor causing a write failure. .  It's SMART info looks OK.

My main concern right now is not to do any more damage than necessary. If I try and replace the parity with one emulated data disk, what will happen?

2 minutes ago, Squamsh said:

If I try and replace the parity with one emulated data disk, what will happen?

 

You would lose the contents of the emulated drive.   

 

That was why I was wondering if the physical disk4 drive has failed as if not you might be able to get at its contents by mounting it under Unassigned Devices assuming that a failed rebuild attempt has not damaged the file system beyond repair.

1 hour ago, itimpi said:

 

You would lose the contents of the emulated drive.   

 

That was why I was wondering if the physical disk4 drive has failed as if not you might be able to get at its contents by mounting it under Unassigned Devices assuming that a failed rebuild attempt has not damaged the file system beyond repair.

Are there any tests to do or is it a case of try and find out?

 

What would be a good order of operations here. I'm thinking something along the lines of

1. Replace parity

2. Remove data disk 4 from array

3. Start parity sync

4. See if I can find some data from disk 4 under unassigned devices

 

Is unassigned devices part of base unraid or do I need the plugin? Can't find it under community apps

If you build parity without disk4 then Unraid will have to clear disk4 when you reassign it. 

 

Do you have backups of anything important and irreplaceable? 

2 hours ago, Squamsh said:

If I try and replace the parity with one emulated data disk, what will happen?

It wouldn't let you unless you New Config (reset disk assignments) which would mean disk4 would no longer be emulated. 

2 hours ago, itimpi said:

assuming that a failed rebuild attempt has not damaged the file system beyond repair.

For future reference, never rebuild onto the same disk unless you are reasonably sure all other disks are working well. Keep original disk as is in case of problems. 

29 minutes ago, Squamsh said:

4. See if I can find some data from disk 4 under unassigned devices

 

Is unassigned devices part of base unraid or do I need the plugin? Can't find it under community apps

 

It is still there when I checked about five minutes ago!  Open up the APPS tab, watching the screen, start typing  "unassigned".  After about three letters, the dropdown box should contain the three plugins that are the total unassigned package.  

9 minutes ago, trurl said:

It wouldn't let you unless you New Config (reset disk assignments) which would mean disk4 would no longer be emulated. 

2 hours ago, itimpi said:

assuming that a failed rebuild attempt has not damaged the file system beyond repair.

For future reference, never rebuild onto the same disk unless you are reasonably sure all other disks are working well. Keep original disk as is in case of problems. 

 

This is great advice.  I have a cold spare ready to go at all times.  (Precleared for both new and recycled older disks.)  If a disk is ever disabled-- usually, these are caused by a failure to write --, I check to see if there is some reason.  If I don't find anything, I remove and set aside the disabled disk.  I install  and assign the cold spare in its place.  Unraid will then rebuild that disk using parity.  I hang onto that removed disk until I have a chance to check that the data is there and everything is working.  I then preclear that removed disk to see if it has any problems at all.  If the preclear cycle report and the SMART report are completely free of any problems, it will become the new cold spare.  (If there is nothing in the SMART Data to indicate that the disk has a problem, I have never had a disk fail to pass a preclear cycle!)

 

Granted, there is the expense of having that cold spare but I feel that I would probably have buy it anyway at some point in the future.  

2 hours ago, trurl said:

If you build parity without disk4 then Unraid will have to clear disk4 when you reassign it. 

 

Do you have backups of anything important and irreplaceable? 

It wouldn't let you unless you New Config (reset disk assignments) which would mean disk4 would no longer be emulated. 

 

Unfortunately I don't really have any backup for that data. Is it possible to remove disk 4 from the array and somehow keep the data on the disk (assuming the data is still intact of course)? To maybe try and extract it with unassigned devices or something similar.

 

Is the data on disk 1-3 safe if I would try and rebuild the raid (removing disk 4)?

How would I best get the array up and running again? Since I can't start it in its current state (too many wrong and/or disks missing). Is it possible to somehow remove disk 4 from the array and get it started from there (and not destroying data on disk 1-3). Disk 4 might be lost but maybe I can try and recover the data from it elsewhere.

You can use the New Config tool and assign disks 1-3 plus the new parity you want to use.  To avoid mistakes I would recommend using the option to keep current assignments and then return to the main tab to make corrections as needed.  Then when you start the array the contents of disks 1-3 will be left intact and parity will be rebuilt based on the contents of disks 1-3.

After you get your array stable again with new parity and disks 1-3, then we can see about recovering data from disk4.

 

Don't assign disk4