IBM serveraid and linux

[Lennart Sorensen]

On Wed, Dec 01, 2004 at 04:25:27AM -0500, Jason Shein wrote:
> Does anyone on this list have any experience with IBM serveraid?
> 
> I have 2 identical Netfinity 5600's setup with RAID 5 across 5 18gb drives, 3 
> live, 2 hot spares. I am constantly having drives go DDD ( amber light ). 
> Randomly. Sometimes it takes 2 weeks running 24/7. Sometimes I will get 2 
> drives on the same day. I have the latest bios and drive updates installed. I 
> have ran all the diagnostics on the IBM serveraid bootable cd. All tests 
> fine. I have tried Fedora, Debian sarge, SuSE 9.1 and 9.0, Now running 
> Whitebox ( because of the IBM support for RHEL3 ). All do the same thing. 
> When these systems were at their original location they ran flawlessly - 
> running unix.
> 
> I will be reinstalling these sytems, as in production I will be running gentoo 
> 2004.3, but I am considering installing freeBSD to try and weed out if it is 
> a linux / unix issue.
> 
> Anyone have any suggestions here ?

Sounds like a hardware problem.  I have had similar problems caused by
both a defective serveraid card (took months for IBM to admit that yeah
the card might be faulty) and from a hotswap backplane with flacky
connectors.

It is VERY unlikely to have anything to do with the OS since all the
handling of bad disks and such is entirely part of the serveraid
firmware (are you running the latest?) and the drives and backplane.

So try swapping out the serveraid card or backplane if possible and see
if it helps.

Oh and is your power supply large enough to run that many disks?  Some
of the IBM boxes come with one power supply and room for 3 and actually
require 2 or 3 installed to run more than a certain amount of hardware.

Lennart Sorensen
--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml