Re: Using RMAN to change DB_BLOCK_SIZE
- From: Greg Rahn <greg@xxxxxxxxxxxxxxxxxx>
- To: Haroon Qureshi <haroon.qureshi@xxxxxxxxx>
- Date: Fri, 22 Oct 2010 17:54:31 -0700
A 30% improvement in what operation? Data load? Query time? Both?
What do you attribute the 30% to? Simply changing the block size
or something else (data file placement, etc)?
What does the performance profile (AWR/ASH top 5) look like for each
operation (16k vs 32k).
The more details you can give the better.
On Fri, Oct 22, 2010 at 5:20 PM, Haroon Qureshi <haroon.qureshi@xxxxxxxxx>wrote:
> Thank you for the tips. My testing included creating a 32k
> tablespace. I ran some tests there and found that we had about a 30%
> improvement with the 32k block size over 16k. This is a DW and we
> want to increase our throughput and performance for our ETL loads. We
> also are moving to ASM.
>
--
Regards,
Greg Rahn
http://structureddata.org
Other related posts: