[SOLVED] Disk removal unmountable after rebuild

[SpyisSandvich]

I have my server outfitted with several disks of varying size, and two 8TB parity disks.

 

Recently, one of my data drives failed, and I bought a replacement.  When I went to put it in, I also inadvertently took out another data disk that I previously assumed wasn't being used in the array, and didn't realize that until the array was rebuilding.  I worried since that left the array unprotected in case another drive failed, but I let the rebuild complete.

 

Now that it's done, I went back and re-added the drive I accidentally pulled out to rebuild that one (it was an empty drive at least), but now the replacement disk is showing as unmountable.

 

Another detail I noticed was that, when the failed disk was rebuilding, it hid a bunch of the shares that existed only on that drive from the Shares tab.  They reappeared after the disk reconstruction.  Not sure if this is relevant.

 

Did I lose that disk?  Or is it possible to salvage it?

Edited May 15, 2020 by SpyisSandvich

[JorgeB]

Please post the diagnostics: Tools -> Diagnostics

[SpyisSandvich]

ca-server-diagnostics-20200515-1019.zip

[JorgeB]

Check filesystem on disk2:

https://wiki.unraid.net/Check_Disk_Filesystems#Checking_and_fixing_drives_in_the_webGui

[SpyisSandvich]

It looks like I need to start in Maintenance mode for that, but it's still running the parity sync on the disk that I accidentally removed.  Should I wait for that to complete?

[JorgeB]

It's up to you, if you cancel the rebuild you'll need to start over, also depends what percentage is done.

[SpyisSandvich]

It wasn't far along, so I stopped it.  Here's the check results with `-nv` flags used:

diskchk.txt

[JorgeB]

Run again without -n or nothing will be done.

 

If it asks for it use -L

[SpyisSandvich]

I'm aware that -n meant that changes wouldn't be committed, I didn't just want to run it blindly.

 

I've run the repair, discarding the log with -L, and these were the results.  Beyond the CRC error near the top of the log, it doesn't look like there were any other issues, and there aren't any further instructions, so I remounted the array, and it looks like everything's in its place and accessible!  It was a 5TB so I can't speak for everything, but the few things I tried are all in working order, and I'm confident enough to resume running the array.

 

Thanks a ton, I thoroughly appreciate the direction.  Second-guessing what I thought the server can take and getting anxious when something doesn't go as expected really puts up blinders to the solution.

[JorgeB]

12 hours ago, SpyisSandvich said:

and it looks like everything's in its place and accessible! 

Should be fine, and if there are any lost/partial files xfs_repair usually places them on a lost+folder found on that disk.