RAID5 Array "Failed logical device segments"

perdrix

Distinguished
Dec 4, 2012
4
0
18,510
I had a RAID 5 array of four 4TB drives on an Adaptec ASR-51245. I installed an extra four drives and added three of them to the array (with the fourth as a hot spare).

The process of reconfiguring has been running now for some days (37% complete) and the RAID card just started screaming at me saying that the array is "Degraded Reconfiguring", while drive 0 (the 1st one) of the array is reporting "Failed Logical Device Segments". The reorganisation of the array appears to be continuing and I've managed to silence the alarm using Adaptec Storage Manager.

I found the following in syslog:

Oct 16 20:59:43 Charon smartd[935]: Device: /dev/sdb [aacraid_disk_00_00_0] [SCSI], SMART Failure: HARDWARE IMPENDING FAILURE TOO MANY BLOCK REASSIGNS

and /usr/StorMan/RaidEvtA.log says:

16 October 2018 22:52:56 BST WRN 301:A01C-S--L00 Charon.home Logical device is degraded: controller 1, logical device 0 ("Shared").

I suspect that so far the drive is still 100% readable.

I've ordered a replacement drive, but wonder what my best recovery approach is once the reconfiguration completes:


    Assuming the drive is still only in "warning" state, mark it as failed, remove it and replace with the new drive, and let the card rebuild the array, or

    Power down and offline copy the old drive 0 to the new replacement?

    Some other approach


Thanks in advance
David
 
If one disk is down and the raid is still working then it needs to start using the hot spare. Lookup how to do that for your raid controller.

The easiest way and best practice is to restore from another backup if the raid is broken.
Trying to figure out raid failures is pretty difficult and very vendor specific.
The actual data will be useless on each disk. The raid needs to be recovered if there is no backup or it's gone.