Re: optimizer parameters
- From: "ed lewis" <eglewis71@xxxxxxxxx>
- To: <dmann99@xxxxxxxxx>, <oracle-l@xxxxxxxxxxxxx>
- Date: Tue, 26 Apr 2011 06:05:06 -0400
Hi Dave,
Yes, I can trace the query in both environments,
and compare the results.
I also doubt that the vendor will push to get
these parameters implemented, knowing that they will be accountable.
In the meantime, I plan to go forward, and handle each
sql query on an individual basis.
I'm still curious though, if anyone has modified, and implemented
these parameters. If you were installing a new application, would
modifying these parameters be a general practice, even if it was
not a vendor recommendation ?
thanks
ed
----- Original Message -----
From: "David Mann" <dmann99@xxxxxxxxx>
To: <oracle-l@xxxxxxxxxxxxx>
Sent: Monday, April 25, 2011 11:06 AM
Subject: Re: optimizer parameters
>Stephane,
Thanks for you input.
I did change these parameters on the session level,
but it did not change the original plan for this
particular query. So even making these changes,
does not guarantee the desired or expected results.
As been mentioned, the focus should be on the
actual query.
Also, this particular query takes 14 seconds on prod,
but 6 seconds in dev. They both generate the same plan.
The environments also differ in data volume and user
activity. The vendor is questioning why the times are
not comparable.
So the plans are the same - I would let the vendor know what you found
- especially that their recommendation did not solve the issue. I
avoid "Big Knob" tuning on established systems as well. Unless the
client has a way to test the system and determine if any performance
got worse because of the changes.
In the meantime anything stopping you from tracing execution (with
WAIT info captured) on Dev and Prod to see differences in what the
query is spending its time on?
-Dave
--
Dave Mann
www.brainio.us
www.ba6.us - Database Stuff - http://www.ba6.us/rss.xml
--
//www.freelists.org/webpage/oracle-l
--
//www.freelists.org/webpage/oracle-l
Other related posts:
