Re: log file sync

On Thu, 12 Oct 2006, Alex Gorbachev wrote:

> Maybe I missed something but, perhaps, you are not scoping your data
> properly in time and/or sessions. If you don't have many XCTENDs than
> you shouldn't have many log file sync waits either.
>
> You said "once the big chunk of processing starts...". In this case I
> assume that this chunk is your business critical part you are keen to
> optimize. Thus, you should limit analysis only to this part.

hi - thanks for the advice.

I have set a trigger to activate a SQL trace when a specific user logs in (this app truly does use only one login, I verified) and I also started a SQL trace on active sessions before the application was started. There are only two active sessions during the run of the application. When I review the tracefile I see quite a lot of activity of the 30 different cursors it opens but no XCTEND until near the very end of the application run. The developer swears that the program issues only one commit while it runs.

I am wondering if there is a commit going on that is not part of the cursor activity but within the same session - however I do not see more than a few XCTEND lines in the traces.

The big question on my mind is whether log file sync is directly attributed to commits or if the volume of inserts is enough to create this. The application does issue many many many inserts.

Thanks!

--
http://www.freelists.org/webpage/oracle-l