Re: Insert contention on RAC
- From: "Rajeev Prabhakar" <rprabha01@xxxxxxxxx>
- To: paul.baumgartel@xxxxxxxxxxxxxxxxx
- Date: Wed, 25 Jun 2008 17:28:10 -0400
Paul,
What is the extent size for the table which is experiencing such waits ?
Perhaps, small extent size is the underlying reason ?
Rajeev
On Wed, Jun 25, 2008 at 5:15 PM, Baumgartel, Paul
<paul.baumgartel@xxxxxxxxxxxxxxxxx> wrote:
> Tim,
>
> No indexes at all; they're built after data population.
>
> I also didn't mention that each run starts with a newly-created table,
> thus the heavy HWM contention.
>
> Roman Podshivalov had a promising idea: insert junk data beforehand to
> raise the HWM, then delete it and run the real thing. There are several
> million rows, so a DELETE might be slow enough to erase any time savings
> in the run itself. I thought of TRUNCATE...REUSE STORAGE, but the docs
> aren't clear about HWM; they say HWM is reset if there's more than one
> freelist, but don't mention what happens if there's only one freelist.
> Guess I'll just have to keep testing.
>
> Thanks to everyone for their helpful suggestions.
>
>
> Paul Baumgartel
> CREDIT SUISSE
> Information Technology
> Prime Services Databases Americas
> One Madison Avenue
> New York, NY 10010
> USA
> Phone 212.538.1143
> paul.baumgartel@xxxxxxxxxxxxxxxxx
> www.credit-suisse.com
--
//www.freelists.org/webpage/oracle-l
Other related posts:
