Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> Re: Stats and behavior weirdness

Re: Stats and behavior weirdness

From: Kevin Lidh <kevin.lidh_at_gmail.com>
Date: Mon, 6 Mar 2006 13:18:05 -0700
Message-ID: <cb4807f0603061218p55972230ybb6039a9ad59019c@mail.gmail.com>


Wow, thank you Juan for those. They are explicitly set by the CRM application to caching=0 and adj=1. When I changed them to 90 and 10 respectively, I got the same explain plan as on my Linux box and production (and another instance tested later). Since I don't have any control over the CRM application and there's a support issue when you make changes, I can "fix" this with a stored outline. But I still don't understand why the development database uses the same information as my Linux copy (as it is just an export) and comes up with a different plan AND why it performs so poorly. There is something out there that these values counter. But this definitely gives me a direction to investigate.

On 3/6/06, Juan Carlos Reyes Pacheco <juancarlosreyesp_at_gmail.com> wrote:
>
> Hi Kevin, why don't you try to set this parameters in one session
> OPTIMIZER_INDEX_COST_ADJ = 10
> OPTIMIZER_INDEX_CACHING = 90
>
> When suddenly, the performance moves from seconds to almost an hours,
> without good reasons, I solve this settings this parameters, when the
> code can be improved.
>

--
http://www.freelists.org/webpage/oracle-l
Received on Mon Mar 06 2006 - 14:18:05 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US