Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Mailing Lists -> Oracle-L -> RE: performance of redo with CIO option on AIX JFS2 filesystem.

RE: performance of redo with CIO option on AIX JFS2 filesystem.

From: Allen, Brandon <>
Date: Tue, 14 Feb 2006 09:26:26 -0700
Message-ID: <04DDF147ED3A0D42B48A48A18D574C4503D40232@NT15.oneneck.corp>

I have recently been working on implementing a similar system and have had good results so far, just a few things to consider:

  1. Depending on your version, you may need to set filesystemio_options=setall (I think you have to in 9i, but in 10g it might not matter)
  2. Make sure you use the chfs command (or whatever it is, I'm not sure) to make the -o cio mount option persistent through server reboots
  3. Create the filesystems for redo logs with agblksize=512
  4. Plan on increasing your db_cache_size to compensate for the loss of file system buffering at the OS level
  5. Reduce minperm, maxperm & maxclient and set lock_sga=true

Search google and metalink for "AIX CIO" and you'll get a lot of good hits.


-----Original Message-----

[]On Behalf Of The Human Fly Sent: Tuesday, February 14, 2006 2:52 AM To: oracle-l
Subject: performance of redo with CIO option on AIX JFS2 filesystem.

Hello list,

Today we have a light smile on our face because we were struggling with log file sync event on one of our business critical database. We opened a TAR and implemented everything they have suggested. Yesterday, we made litte changes, like, created a new mountpoint enabling CIO (concurrent I/O) on AIX 5.3 using JFS2 filesystem, and move our redo groups to this new mount point also increased the size of redo groups from 512M to 1gb.
When we look at the log file sysn event in the statspack, it was just vanished. Earlier, we used to get 40% as log file synce in the top timed events and now it is just 3%. Its really a drastic change in the wait event.
Now, I am planning to do the same for our datafiles, i.e. enabling cio and moving the tablespace to new moutn point. Redo has only inserts, but, for tables, we have read and write, does this can improve the throughput? Will there be any performance gain? Does anyone using this kind of option with JFS2 filesystem?

Thanks for your time.


Best Regards,
Syed Jaffar Hussain
8i,9i & 10g, OCP DBA
Banque Saudi Fransi,
Saudi Arabia

"Winners don't do different things. They do things differently."

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.

-- Received on Tue Feb 14 2006 - 10:26:26 CST

Original text of this message