Re: 11.2.0.3 SQL Profile Confusion
From: Andy Klock <andy_at_oracledepot.com>
Date: Fri, 1 Nov 2013 15:34:42 -0400
Message-ID: <CADo_RaMbXcip_Yrj0ab6RvOWr-kEacdGaE99Sfg+Zqk-3dZNDg_at_mail.gmail.com>
On Fri, Nov 1, 2013 at 2:52 PM, Chris Taylor <christopherdtaylor1994_at_gmail.com> wrote: ...
>
> Also, when I tested it out, elapsed time jumped from 1.8 Seconds to 26.845
> seconds per execution and used 32 parallel processors! (I definitely
> understand this - way too many parallel processes and the overhead involved
> increased the total execution time).
Date: Fri, 1 Nov 2013 15:34:42 -0400
Message-ID: <CADo_RaMbXcip_Yrj0ab6RvOWr-kEacdGaE99Sfg+Zqk-3dZNDg_at_mail.gmail.com>
On Fri, Nov 1, 2013 at 2:52 PM, Chris Taylor <christopherdtaylor1994_at_gmail.com> wrote: ...
>
> Also, when I tested it out, elapsed time jumped from 1.8 Seconds to 26.845
> seconds per execution and used 32 parallel processors! (I definitely
> understand this - way too many parallel processes and the overhead involved
> increased the total execution time).
I don't think a SQL with an elapsed time is necessarily a good candidate for parallel queries.
>
> Anyone want to help me understand why it came up with a Parallel SQL Profile
> recommendation in the first place?
SQL Tuning Advisor was grasping at straws? :)
Probably better just to determine where time is being spent and where it can possibly be tuned. If you want to run it through a tool, how about giving SQLT a go? It provides you with some real data that you can use to come up with a better guess than 96.52% Benefit :)
Andy K
-- http://www.freelists.org/webpage/oracle-lReceived on Fri Nov 01 2013 - 20:34:42 CET