Re: RAID problem

From: Benjamin Goldsteen (ben_at_inka.mssm.edu)
Date: 08/28/03 

Date: 27 Aug 2003 19:19:46 -0700

gheston@hiwaay.net (Gary Heston) wrote in message news:<vko3b1fb02cd3b@corp.supernews.com>...
> In article <b6479101.0308252121.2e0ab419@posting.google.com>,
> A. K. Patnaik <atulpatnaik@yahoo.com> wrote:
> >Hi,
>
> >We have a Fiber RAID (CLARIION Make) connected on SGI ORIGIN200 system
> >with 9+1 disks as a hot spare configured as RAID3. Our problem is the
> >Amber LED is glowing for 2 disks (disk2 & Disk5). In the log file
> >there are soft SCSI error as well as Hard SCSI error and sometimes CRU
> >shutdown errors are there. After recycling the power to RAID it comes
> >to a normal state and data-rebuilding process starts. Immediately
> >Completion of this processes again the disks starts showing faulty.
> >I have changed the default SP, the Fiber Channel controller and the
> >connecting cables, but still the problem persists and goes off after a
> >power recycling to the RAID. Can any body suggest me where could be
> >the problem. Please help me and it will be greatly appreciated.
>
> Drives 2 and 5 are failing and need to be replaced. Now.

I don't want to be the messenger of bad news, but I think it might be
too late. My guess is that there are irrecoverable bad blocks or
something across multiple disks and that the rebuild will fail. I
would backup the RAID before trying the replace/rebuild. If it still
fails, I would backup, replace both drives, and reformat/rebuild the
LUNS, restore data...

Assuming both drives didn't fail at exactly the same time, the
question is why didn't the hot spare come online when the unit
detected one or the other drive failing.

Relevant Pages