Re: Asynchronous Commit in Oracle Database 10g R2

Mladen Gogala wrote:
> It is my own experience. I created a table A(c1 number), executed
> the loop from post above and counted "io_submit" calls.
> Eventually, I counted the following:
>
> grep io_submit /tmp/lgwr.out|wc -l
> 4408
> And user commits (v$sesstat) was 5002. The obvious conclusion is
> that not every commit has caused write request to be generated.

But it doesn't have a regular pattern, does it? I mean: it's not like Oracle is optimizing *every* single turn through the loop. If there truly is an optimization in 9i then it should be a regular, repeatable event, no?

I'm wondering if this is not a simple manifestation of the commit-piggyback that has been around since v6. Ie: if PL/SQL session somehow gets interrupted and there is a context switch, particularly in a multi-cpu environment, there is the possibility that logwriter would wake up, see "two" requests to commit from two sessions in two cpus and say: hang-on, I only need to commit once. Received on Mon Aug 29 2005 - 01:38:22 CDT