Re: shared_pool_size and percentage utilized

From: Martic Zoran <zoran_martic_at_yahoo.com>
Date: Mon, 11 Apr 2005 05:40:58 -0700 (PDT)
Message-ID: <20050411124059.86287.qmail@web52601.mail.yahoo.com>

Sami,

If neither of these things are true then NO:

Is any critical business unit spending considerable CPU time on parsing then on anything else?
Global database view: is your "parse time cpu" huge in regard to overall CPU time?

Maybe some addition to these two rules are possible.

I have the customer who does not want to increase buffer cache more then 4G (even having 96G of memory on the box) while putting shared_pool_size to 1G without any reason. Apps for long running SQL's are not using binds and who cares, while fast SQL's are doing prepared statements with binds anyway.

It is hard to explain to many DBA's, not just developers that 0.01s of parsing while SQL spending more then 1s or sometimes even 1000 of seconds are not worth it tuning.

Regards,
Zoran

Sami Seerangan <dba.orcl_at_gmail.com> wrote:

> Hi :
>
> I did run the following program(from metalink) in 8i
> (2 node OPS)
> during peak hours and it says shared_pool
> utilization is 165%. Do I
> have to increase the shared_pool_size to improve
> the performance?

Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

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

Received on Mon Apr 11 2005 - 08:44:55 CDT