Re: COMMIT NOWAIT Performance Option

Joshua D. Drake wrote:
> Gregory Stark wrote:
> > "Joshua D. Drake" <jd(at)commandprompt(dot)com> writes:
> >
> >>>> On 2/27/07, Josh Berkus <josh(at)agliodbs(dot)com> wrote:
> >>>>> I see no reason to implement it if there is no performance gain.
> >
> >> However, I strongly concur that we need at least some evidence. It could
> >> easily be that a misstep in the code, causes a loop over the wrong set
> >> and all the performance we thought we would get is invalid, not because
> >> of theory or what should happen, but because of actual implementation.
> >
> > It rather sounds like you're asking for a proof that Simon can write bug-free
> > code before you allow him to write any code.
>
> Well wouldn't that be great! :) but no, not quite. I would just like to
> see some metrics showing that it is a benefit. Besides the patch needs
> to work for the metrics to be run.

I don't understand the great demand for metrics at this point. Once the
patch is ready, people can run the patch on their workloads to get
real-world metrics. Metrics are only needed before the patch is
applied, not before it is discussed.

--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://www.enterprisedb.com

+ If your life is a hard drive, Christ can be your backup. +