RE: optimizer_ ???

Please don't infer from my response that I am an amphibious bird of any sort but I might take the following approach even though it is a bit unscientific.

I typically have a program running that snapshots my DB workload (IO, CPU etc...) and waits each hour, this information is easily summarized into daily/hourly breakdowns. Over time assuming a reasonably stable application you get very consistent numbers. Now when you go tweak a parameter that will effect the entire system you should see these numbers change and the change should once again be consistent over time. If the change is good then one could reasonably assume response times in theory would improve (i.e. less physical IO and CPU stays stable and less IO wait).

-----Original Message-----
From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Jonathan Lewis Sent: Wednesday, April 27, 2005 3:27 AM
To: oracle-l_at_freelists.org
Subject: Re: optimizer_ ???

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