Re: 8k block size upgrade concern

"JeffV" <jeffv27982_at_aol.com> wrote in
news:1119977519.564427.326060_at_z14g2000cwz.googlegroups.com:

> Thanks. The upgrade is part of a much larger program for achieving
> performance gains. There has been considerable time spent over six
> months analyzing the "characteristics" of the batch and online
> application and these solutions are a result of that analysis. My
> belief is certainly that we will see a limited a performance gain from
> the upgrade. It is also safe to say I know the upgrade will go on
> without a hitch from a data quality standpoint. My concern is the
> unknown about the possibility of the optimizer choosing poor plans due
> to the switch. I believe the with the right optimizer choices, the
> batch will perform faster than it does now. Other sub-projects in the
> program will have to be put in place to see the full performance
> impact.
> Thanks,
> Jeff
>

What is the expected end run time?
Down from 44 hours to what value?

What percentage of the expected reduction is to come from 8k blocks?

Then I would first do what is expected to make the greatest % improvement. If the actual results differ from expectation, you may wonder what is/was wrong with your anaylsis.

If your "best shot" misses expectations by a wide margin, why would a reasonsible & reasonable person expect better results from the other planned activities.

Please don't confuse me with the facts, my mind is already made up. Received on Tue Jun 28 2005 - 20:53:46 CDT