We had a good experience, but it was during an implementation of a new
server so there wasn't a whole log of pre-CIO data to compare to,
however we did do some before/after testing and had improvement with CIO
but it is critical that you also increase your db_cache_size when you do
this, and this also makes it difficult to compare apples to apples -
because obviously increasing your db_cache_size can improve performance
regardless of CIO. You are basically stealing the cache away from the
OS level where Oracle would have reported incredibly fast physical reads
because they were in fact not physical at all, but were cached by AIX -
so you have to replace that with more cache for Oracle (we increased
db_cache_size from 800MB pre-CIO to 2GB post-CIO)

I have since implemented a 2nd system with CIO and this time didn't even
do a pre-CIO comparison - just went with it from the beginning and
performance has been excellent. Also, be sure to set agblksize=512 for
you redo log filesystems, set filesystemio_options=setall and verify
your filesystems are mounted with o=cio if using 9i.

Regards,
Brandon

-----Original Message-----
From: oracle-l-bounce@(protected)
[mailto:oracle-l-bounce@(protected)
Sent: Wednesday, June 28, 2006 1:07 PM
To: oracle-l@(protected)
Subject: Concurrent I/O (AIX 5.2)

We bit the bullet and implemented concurrent I/O (AIX 5.2 ML6, 9.2.0.5,
JFS2, ESS storage).



Privileged/Confidential Information may be contained in this message or attachments hereto. Please advise immediately if you or your employer do not consent to Internet email for messages of this kind. Opinions, conclusions and other information in this message that do not relate to the official business of this company shall be understood as neither given nor endorsed by it.

--
http://www.freelists.org/webpage/oracle-l