Re: system stats and LIO's
- From: Jared.Still@xxxxxxxxxxx
- To: oracle-l@xxxxxxxxxxxxx
- Date: Thu, 5 Aug 2004 15:21:44 -0700
> Hi,
> I am testing the impact of system stats.
> From reading the list here I know the best goal to follow is to reduce
> LIO's, not exec time, exec path, cost, etc.
> So, how do I determine LIO's for a given statement with and without
stats.
> Thanks
> --
> ..
Though LIO's are important, you need to determine which statements are
worthy of your tuning efforts.
A few things to consider:
* response time
* frequency of execution
* how important the results are to the users
Say you find a statement with a high number of LIO's
( aka a high buffer hit ratio ), that takes 5 minutes to
execute and is run once every day at 3:00 AM.
You may correct the problem and make it fly, completing its
run in 12 seconds. The problem is, no one will care.
You need to work on things that people will notice, the slow
jobs that drive the users crazy. These are the important ones.
Isolate these transactions, find out what makes them slow, fix
the problem if possible, and move on to the next one.
This sounds a lot like a lead in to "Optimizing Oracle Performance"
doesn't it?
"And now, heeeeerrrress's Cary and Jeff, brought to you by HOTSOS.COM,
world leaders in Oracle performance diagnosis and Habenero Salsa".
"Take it away Cary!"
http://www.oreilly.com/catalog/optoraclep/
Jared
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
Other related posts:
