Sigh. So my UnRaid system (running 6.1.9) crashed, requiring me to do a hard reboot*. When I restarted, however, my RAID array seems to have gone. No array devices are listed and all of my disks are listed as unmountable. ARGH!

 

I would restore the config file from backup, but the USB key I kept it on seems to have died. DOUBLE ARGH!

 

My array is built from 13 disks of various sizes, including 3 4TB disks, one of which was the parity disk. But without the backup of the config file, I can't figure out which it is. Obviously, I don't want to start randomly rebuilding things until I figure out where to start.

 

So now I have no idea which disk was the parity, etc, and don't want to do anything that might compromise the data. What should I be doing in this situation to try and recover the data?

 

 

* note this was a tital crash: no response at the console, no response to net traffic, etc.

Go to Tools - Diagnostics and post your complete diagnostics zip.

Here is the diagnostics.

bernard-diagnostics-20161213-1509.zip

Your diagnostics seem to indicate that you are randomly assigning drives trying to figure them out. STOP!

 

If you don't know which disk is parity, you must avoid assigning any disk to parity, since if you manage to start the array with a data disk in the parity slot, that data will be overwritten with parity and the data from that disk would be lost.

 

Do you know what filesystem was on the drives?

Hm, I wasn't assigning stuff randomly, but I will stop fiddling. All of the disks were XFS, and two SSD cache drives (BTFS?).

Maybe not randomly, but there was some assignments and reassignments going on.

 

Assign all but the SSDs as data disks. DO NOT assign any disk to the parity slot. Then post a screenshot of Main - Array Devices.

Something strange is happening here, don't remember seeing it before, all disks were mounted individually before the array start:

Bernard kernel: XFS (sdf1): Mounting V5 Filesystem
Dec 13 14:31:58 Bernard kernel: XFS (sdf1): Starting recovery (logdev: internal)
Dec 13 14:31:58 Bernard kernel: XFS (sdf1): Ending recovery (logdev: internal)
Dec 13 14:32:03 Bernard kernel: XFS (sdh1): Mounting V5 Filesystem
Dec 13 14:32:04 Bernard kernel: XFS (sdh1): Starting recovery (logdev: internal)
Dec 13 14:32:04 Bernard kernel: XFS (sdh1): Ending recovery (logdev: internal)
Dec 13 14:32:12 Bernard kernel: XFS (sdi1): Mounting V5 Filesystem
Dec 13 14:32:13 Bernard kernel: XFS (sdi1): Starting recovery (logdev: internal)
Dec 13 14:32:13 Bernard kernel: XFS (sdi1): Ending recovery (logdev: internal)
Dec 13 14:32:21 Bernard kernel: XFS (sdj1): Mounting V5 Filesystem
Dec 13 14:32:21 Bernard kernel: XFS (sdj1): Ending clean mount
Dec 13 14:32:48 Bernard kernel: XFS (sdk1): Mounting V5 Filesystem
Dec 13 14:32:48 Bernard kernel: XFS (sdk1): Ending clean mount
Dec 13 14:32:56 Bernard kernel: XFS (sdl1): Mounting V5 Filesystem
Dec 13 14:32:56 Bernard kernel: XFS (sdl1): Ending clean mount
Dec 13 14:33:01 Bernard kernel: XFS (sdm1): Mounting V5 Filesystem
Dec 13 14:33:01 Bernard kernel: XFS (sdm1): Starting recovery (logdev: internal)
Dec 13 14:33:02 Bernard kernel: XFS (sdm1): Ending recovery (logdev: internal)
Dec 13 14:33:11 Bernard kernel: XFS (sdn1): Mounting V5 Filesystem
Dec 13 14:33:11 Bernard kernel: XFS (sdn1): Ending clean mount
Dec 13 14:33:16 Bernard kernel: XFS (sdo1): Mounting V5 Filesystem
Dec 13 14:33:17 Bernard kernel: XFS (sdo1): Ending clean mount
Dec 13 14:33:27 Bernard kernel: XFS (sdp1): Mounting V5 Filesystem
Dec 13 14:33:28 Bernard kernel: XFS (sdp1): Starting recovery (logdev: internal)
Dec 13 14:33:28 Bernard kernel: XFS (sdp1): Ending recovery (logdev: internal)
Dec 13 14:33:38 Bernard kernel: XFS (sdq1): Mounting V5 Filesystem
Dec 13 14:33:38 Bernard kernel: XFS (sdq1): Ending clean mount

 

So when you tried to start the array they didn't mount because of the duplicate UUID, can't mount the same UUID twice.

 

I think (sde) ST4000DM000-1F2168_Z301HSRK is your parity because it's the only one without a duplicate UUID, unRAID didn't found a filesystem, normal for the parity disk.

 

I see you have the Unassigned Devices plugin installed, reboot and before starting the array (and all disks unassigned) look at the unassigned devices, parity should be the only one without a mount option, if it's the only one (and all the other disks show xfs) reassign all disks including parity, check the box "parity is already valid" and start the array.

Okay, here is the array devices screen. Several of the disks get green blobs when I added them, which I assume means that they are valid array devices. Others did not, though. I am guessing that one of these was the parity disk. Are the others likely scrambled?

 

 

 

(Oh, and there was an additional 1TB drive that wasn't part of the array. That was assigned directly to a VM as it was a scratch disk for live TV streaming).

YAY! Okay, so you were right: I rebooted, looked for the one with no file system, then added that as the parity disk and the rest as data. The array started up okay, and I seem to be able to read from it okay.

 

THANK YOU! I thought I was screwed for a bit there....