Re: Timestamps in trace files (and other trace file oddities)

From: Daniel Fink <daniel.fink_at_sun.com>
Date: Thu, 24 Jul 2003 16:49:04 -0600
Message-Id: <26007.339494@fatcity.com>

This is a multi-part message in MIME format.

--------------37BF07393403ED9FD641B4C9
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

10046 Trace files are a peak into how an application and Oracle operate. In looking at one generated by a web server, I identified that the application (user) was sorting to a temporary tablespace that was dictionary/permanent without at the data dictionary. I saw a query performing DML against fet$, uet$ and seg$. You can see every step that the application performs, how Oracle responds. You can even see what scripts are called from a sql*plus session! COOL! (then again, I like to read formatted block dumps to determine the undo chain!)

Another significant area of processing can be sorting. I don't know if sort processing is counted along with LIOs or PIOs.

Dan

Freeman Robert - IL wrote:
>
> If you have a 10046 trace file, and the means to process/analyze it, then it
> will tell you what you need to know. If it's IO, you will see IO related
> waits. If they are not there, and no other significant wait events, then
> it's all processing time, which probably equaits to LIO's. Thats my take.
>
> Love 10046's.... just had a project where they really helped me figure out
> what was going on.
>
> RF

--------------37BF07393403ED9FD641B4C9
Content-Type: text/x-vcard; charset=us-ascii;  name="daniel.fink.vcf"

Content-Transfer-Encoding: 7bit
Content-Description: Card for Daniel Fink
Content-Disposition: attachment;

 filename="daniel.fink.vcf"

begin:vcard
n:Fink;Daniel
tel;cell:303.808.3282
tel;work:303.272.3225
x-mozilla-html:TRUE
adr:;;;;;;
version:2.1
email;internet:daniel.fink_at_sun.com
title:DB Services Lead
x-mozilla-cpt:;-4832
fn:Daniel Fink
end:vcard

--------------37BF07393403ED9FD641B4C9-- Received on Thu Jul 24 2003 - 17:49:04 CDT