Re: Correct uniform size for temporary LMT ?

> Tablespaces don't cost more. I can not imaging using less than three
> tablespaces for data and three for indexes.
>
> Then size them, after examining your segments, to something like this:
>
> 128K, 5M, 25M
>
> so that you have small for lookups, medium where appropriate, and large
> for the largest: Usually I use more depending on my analysis and getting
> the best possible usage of my disk space.
>
> Daniel Morgan

Thanks Daniel, but how about for my temporary segments ?

In V$SYSSSTAT: sorts (disk)/sorts (memory) = 0.2% so I dont think I have to increase the sort_area_size parameter.

One question, If the average sort on disk is lets say 180Kb, then i suppose I should choose an uniform extent size closest to 180Kb (sort_area_size * n) ? Or am I way off here ?

How can I get statistics about an average sort on disk by the way ? I´m not so familiar with the Statspack.

Thanks for any comments.

/Christian Received on Thu Feb 28 2002 - 02:03:58 CST