Re: testing memory parameters

On Mar 17, 4:30 am, Serge Rielau <srie..._at_ca.ibm.com> wrote:
> > Below what value is BCHR "bad"?
>
> In DB2 I treat any OLTP app with less than 90% hit ratio as troubled.
> Anything less than 80% I'd call a mess requiring some serious explanation.
>
> > Above what value is BCHR "good"?
>
> There is no good (we agree here). There only is "not obviously bad".
> Just like a car whose engine check light is off isn't by definition a
> good car.
>
> > What are the exceptions to the two answer to above questions?
>
> Warehousing.
>
> > What are the exceptions to the exceptions?
>
> Too many $ spent on memory.
>
> You never can have clear rules. You only can have indicators. The
> ability to separate an indicator from a rule is part of teh job.
> I don't think we are on opposing ends here. I just find the word
> "useless" swinging the pendulum way too far to the other end.

There is a book called "Tales From The Oak Table" which has some interesting technical history about Oracle, as well as one of the more cogent descriptions of why a low hit ratio may perform better than a high one. If you were working primarily with Oracle I'd probably suggest other things by the writers in that book, but it is an excellent combination of entertaining stories and enough technical depth to appropriately answer this basic question.

The general mistake made with this particular hit ratio, and the one you appear to be making, is that the ratio is an indicator of sql efficiency. It is not. In fact, the general case is that inefficient sql makes the the ratio appear good. Don't know if this link will work, but give it a try:

http://books.google.com/books?id=mvJW6t7mYU0C&pg=PA18&lpg=PA18&dq=lio+pio+cary+milsap&source=web&ots=p3sqWxhdma&sig=ba-LSDQFWvLTrVt0pNviemqZrjQ&hl=en#PPA5,M1

jg

--
@home.com is bogus.
12 people were fired and 13 people disciplined (including doctors) for
checking out Britney's medical records.