UPDATE: SUMMARY: High IO problem after installing 8_Recommended

From: Eric Paul <epaul_at_profitlogic.com>
Date: Mon Mar 29 2004 - 17:15:50 EST
Sun has just released patch 117000-03 which fixes this issue.  This patch is
also included in the 2004/03/26 issue of the Recommended Patch Cluster.

I installed it today, and it does fix that problem.

NB, there is a known bug with this patch, as it does not include the modified
.h files to compile against, so if you are writing code that looks at how much
IO is going on in the kernel, you'll still be in trouble until -04 is released
in a few weeks.  There is a T patch available if you have support and want to
assume the risks that go with such things.

HTH!

Eric

-----Original Message-----
From: Eric Paul [mailto:epaul@profitlogic.com]
Sent: Thursday, March 11, 2004 10:27 AM
To: Randy Romero; Sun Managers List
Subject: RE: SUMMARY: High IO problem after installing 8_Recommended


I know this is somewhat against etiquette, but the info is important enough
that I hope no one minds me adding info to a Summary that is not mine.

I also experienced this problem (although I didn't realize it until I read
this summary this morning.)  I just got off the phone with Sun and received
the following information:

Yes, it is a known bug with 108528-29 and Veritas Foundation Suite.  (Someone
decided to change the format of a data structure, and add new stats to the
MIDDLE instead of the END, so Veritas writes it's I/O performance data
directly to Kernel-space and now those locations are wrong, so it looks to
the
Kernel like there are BILLIONS AND BILLIONS OF WRITES when in fact there are
not.)  Sun assures me this problem is completely cosmetic and there is no
reason to panic.  A patch will be released (to fix the patch :P) at the end
of
the month.  If you have a support contract, you can get a Test (T) patch for
this if you want.

The Bug ID for this issue is 4978228.

Sun hopes to have a real patch for this issue at the end of the month.  It
will most likely be released as a supporting patch (117000-03) as opposed to
a
patch to the Kernel itself.

If you are experiencing the problem now, as Randy Romero pointed out, you can
back out to -27 of the Kernel (and back out all dependant patches) to make
this issue go away.

HTH!

Eric

-----Original Message-----
From: Randy Romero [mailto:rromero@fury.to]
Sent: Wednesday, March 10, 2004 5:51 PM
To: Sun Managers List
Subject: SUMMARY: High IO problem after installing 8_Recommended


Thanks to everyone who responded--too many to list.  The solution was to
back down 2 kernel revisions to 108528-27.  In doing this though, I had to
remove 24 other patches that the cluster installed before I could remove
108528-29.  This was a pain, but after removing the patches, installing
the new kernel, and rebooting the problem was resolved.  Thanks again.




Randy Romero


In a world without walls and fences, who needs Gates and Windows?
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
Received on Mon Mar 29 17:22:28 2004

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:43:30 EST