Re: Journaling without commitment control (WAS iSeries and SSD drives)

[Vernon Hamberg <vhamberg@xxxxxxxxxxxxxxx>]

Yep, a former IBMer, Rick Turner, has an article suggesting this 
technique. It gives almost the same benefit as getting the journaling 
performance OS option (42 or 43 or like that). There's a very small 
window where cached data might be lost in the event of an outage, as I 
recall. But someone else please check that.

Here's a link to where y'all can get the article -

http://tinyurl.com/batch-performance

HTH
Vern

On 1/18/2013 9:00 AM, Charles Wilt wrote:
> Dieter,
>
> Journaling without commitment control has a significant impact on batch
> jobs doing writes...the more writes, the more impact.
>
> About the only scenario where it would NOT have an impact is if all the
> tables you're writing to have FRCRATIO(1) anyway...
>
> Charles
>
> On Fri, Jan 18, 2013 at 2:06 AM, D*B <dieter.bender@xxxxxxxxxxxx> wrote:
>
> > on a balanced system journaling with or without commitment controll
> > doesn't have measurable performance effects, as long as you don't need
> > maximum performance for bulk operations writing some million of records
> > with one SQL statement.
> >
> > abusing commitment controll for multi transactional workload, you have to
> > have in mind lock contentions caused by record locks lasting longer as some
> > miliseconds, that's not only a performance issue, other jobs, waiting for a
> > locked record might break down or not work as designed, if they are not
> > transaction safe (don't use commitment controll).
> >
> > D*B
> > --
> > This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
> > To post a message email: MIDRANGE-L@xxxxxxxxxxxx
> > To subscribe, unsubscribe, or change list options,
> > visit: http://lists.midrange.com/mailman/listinfo/midrange-l
> > or email: MIDRANGE-L-request@xxxxxxxxxxxx
> > Before posting, please take a moment to review the archives
> > at http://archive.midrange.com/midrange-l.