Re: Global Cache and Enqueue Services statistics
- From: Tim Gorman <tim@xxxxxxxxx>
- To: ORACLE-L <oracle-l@xxxxxxxxxxxxx>
- Date: Tue, 31 Jul 2012 12:28:20 -0600
How did the load test do when you ran it from just one instance started with
CLUSTER_DATABASE=FALSE?
On 7/31/2012 12:09 PM, Hameed, Amir wrote:
>
> Thank you Tim and Gaja for your feedback. It is much appreciated.
> Here is some more information and the reasoning behind my mentioning of
> FREELSITS and FREELIST GROUPS. Based on the top waits in the AWR report,
> I queries DBA_HIST_ACTIVE_SESS_HISTORY and was able to identified
> statements that were waiting on these events. There are a total of 12-14
> statements that account for around 90-95% of these waits. These
> statements are SELECT FOR UPDATE, UPDATE, DELETE AND INSERT. These
> include two INSERT statements and one of them is parallelized with over
> twenty sessions inserting into it in parallel. This was the primary
> reason for looking into FREELISTS and FREELIST GROUPS for a few tables
> that were involved in the INSERT statements. As I had mentioned in my
> previous note, this system was upgraded from 11.0.3 to 11.5.x several
> years ago and is therefore not configured with ASSM. Additionally, it is
> my observation that prior to EBS 11i, Oracle did not have a standard on
> defining FREELIST GROUPS on the base Oracle tables and most of the
> tables that I have looked at have these values set to 1. Starting with
> 11i fresh installs (installers prior to 11.5.10), tables started to get
> created with FLG of 4; at least this is what I have observed. With
> 11.5.10.2 installers, Oracle uses ASSM.
--
//www.freelists.org/webpage/oracle-l
Other related posts:
