Re: Odd maximum elapsed event times in raw 10046 trace

• Post for FREE via your newsreader at post.usenet.com ****

Hi Matt

> I've noted some peculiarities in the trace
>
> 1) There is a single large SQL*Net message from client elapsed time
> that may be skewing my analysis.
> 2) There is a single large db file scattered read elapsed times that
> seems odd (as it falls inbetween a number of "standard" sized elapsed
> values)
>
> As the application is a batch processor (third party). I am confused
> by the large value from 1) above. could this simply be the time it
> took me to switch off tracing?

Since a COMMIT follow the 'SQL*Net message from client' the batch was either idle or doing some processing without db access.

> Can anyone please explain 2) above? Does this indicate a potential
> disk problem?

If it happen only once, I'll not take care of it, i.e. I'll "comment" the line from the tracefile for the analysis. If it occurs again and again, there's a problem somewhere. Of course the disk it's the first place to look at...

Chris

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

• Usenet.com - The #1 Usenet Newsgroup Service on The Planet! *** http://www.usenet.com Unlimited Download - 19 Seperate Servers - 90,000 groups - Uncensored -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=