Newly installed RAID drive fails during rebuild

[JamesJ_1]

Hello,

I have an HP Netraid LC3 server that gave the alarm that a drive had failed. We replaced the drive, but it still gave the alarm. The drive failed again after we tried to Rebuild manually. I have included the log history below.

Log Started.
- Wed Mar 18 16:57:40 2009
Rebuild Started on CHNL 0 TARG 0. - Wed Mar 18 17:05:43 2009
Error on Rebuilding CHNL 0 TARG 0 - Wed Mar 18 17:28:22 2009
Notify message : DRIVE STATE changed in Ch 0 ID 0 to FAILED - Wed Mar 18 17:28:47 2009
NOTIFY:Check Condition on Ch 0 ID 1 with the following sense key - Wed Mar 18 17:29:13 2009
f0 00 03 01 6c d3 20
18 00 00 00 00 11 00
NOTIFY:Check Condition on Ch 0 ID 1 with the following sense key - Wed Mar 18 17:29:13 2009
f0 00 03 01 6c d3 20
18 00 00 00 00 11 00
Log Started.
- Thu Mar 19 09:42:32 2009
NOTIFY:Check Condition on Ch 0 ID 1 with the following sense key - Thu Mar 19 09:42:32 2009
f0 00 03 01 6c d3 20
18 00 00 00 00 11 00
NOTIFY:Check Condition on Ch 0 ID 1 with the following sense key - Thu Mar 19 09:42:32 2009
f0 00 03 01 6c d3 20
18 00 00 00 00 11 00
Please let me know of any suggestions that might fix this problem.

Thanks

[kris rombauts]

James,

can you tell us what disk failed (SCSI ID 0) ?

What the log shows is that the array is rebuilding onto disk ID=0, so i guess this is the one that was failed and which has been replaced, however after 23 minutes it fails.

Rebuild Started on CHNL 0 TARG 0. - Wed Mar 18 17:05:43 2009
Error on Rebuilding CHNL 0 TARG 0 - Wed Mar 18 17:28:22 2009


However the disk with SCSI id=1 has a problem with unrecoverable read errors (03 11) on the media as hown by the SCSI sense keys you provided... this means that, eventhough the new disk (ID=0) is likely good, the controller cannot rebuild the data because it encounters read errors from the source disk, so it cannot re-write that part of the data onto the new target disk (id=0).

It is possible that the part that cannot be read is not in real user data, so you can try and take a full backup now to safeguard the data but you will have to re-initialize the array and re-install/re-image the OS to resolve this problem, there is no way you can fix this because one disk is new/empty and the other has some problems on the media.

There is a Netraid monitor program that periodically (weekly) does a compete check
of the array, this is usefull to detect bad spots and fix them before a disk is bad.
It is available on hp.com, i advise , if not already, to install it for the future, i believe it is Windows only. Installing it now makes no sense because it only runs when a array is redundant of course.


HTH
Kris

[JamesJ_1]

Kris,

Thanks for the info.

James