Re: SSA Array Invalid Data Strips

From: support (support@xiatecs.net)
Date: Thu Dec 12 2002 - 22:04:20 EST


Lewis,

We had the same problem few weeks ago on a production system. It was also
after loosing a disk and replace in an SSA RAID5 array. We practically did
the same thing you did. We applied the latest Microcode to the Adapter and
disk, and on top of that we applied ML10. But nothing solved out problem.
IBM faxed me an internal document that CE uses to analyze the SSA adapter
error that you get hourly in the error log. tracing the error all the way to
the invalid strip error. The solution in that document was to
rebuild the SSA RAID run certify on each of the disk to make sure that there
are no bad block and recreate the Array.

What type of SSA disks/ adapter do you have with there Microcode level
please. To me, it looks like this problem is related to a certain type of
adapters or disks as we did not have such a problem with other servers.

Following is the possible causes of the SSA error invalid data strip:
-An unreadable data sector existed on a disk drive when the array was
created
-An unreadable data sector was detected during a rebuilding operation on an
array.
- The SSA adapter was reset during an attempt to recover a failed disk
drive.

XiaTecs Support Team
support@xiatecs.net
www.xiatecs.net

-----Original Message-----
From: IBM AIX Discussion List [mailto:aix-l@Princeton.EDU]On Behalf Of
Lewis, Christina
Sent: Wednesday, December 11, 2002 3:42 PM
To: aix-l@Princeton.EDU
Subject: SSA Array Invalid Data Strips

Hi!!!

I am inquiring to find out if anyone out there has experienced Invalid Data
Strips on their SSA Arrays? If so, how often?

We have invalid data strips occuring every few months and everytime it is
addressed with IBM we are being told to upgrade to the latest level of
microcode. We have upgraded in the past and still experience invalid data
strips. When an invalid data stip does occur, it is always when a disk is
lost - but IBM says that losing a disk does not cause an invalid data strip
and they are saying this is just coincidence.

I would like to know if anyone else is experiencing similar problems?
I am having a hard time accepting their suggestions to upgrade to the new
microcode everytime this happens.

Tks
Tina Lewis
AIX System Administrator



This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 22:16:24 EDT