Re: Cache Hit Ratio from system views

On Sep 8, 4:18 pm, "Bob Jones" <em..._at_me.not> wrote:
> "DA Morgan" <damor..._at_psoug.org> wrote in message
>
> news:1189210426.519912_at_bubbleator.drizzle.com...
>
>
>
>
>
> > Bob Jones wrote:
>
> >>> To quote the Oracle docs:
> >>> "A low cache hit ratio does not imply that increasing the size of the
> >>> cache would be beneficial for performance. A good cache hit ratio could
> >>> wrongly indicate that the cache is adequately sized for the workload."
>
> >> Finally, something useful, but quoted from the doc. Notice the words like
> >> "could" or "would"?
> >> The second sentence is a little vague. What does "good" mean? Is my cache
> >> still too small, if BCHR is 99%?
>
> > You have been challenged repeatedly to put up a test case that supports
> > your position. Either you don't have SQL*Plus or don't know how to use
> > it? I wonder which?
>
> Now you just don't know what you are talking about.
>

Then put the question to rest and PROVE your contentions. Oh, that's right, you can't formulate a concrete example where the BCHR is a useful, reliable metric of database and buffer cache health and performance.

> > You are in no position to criticize anyone or anything. Especially
> > Oracle docs for vagueness. So far the only thing you've done with
> > clarity is define your lack of expertise on the subject.
>
> If the word "good" is clear to you, you should not even attempt to judge
> anything.

The same can be said of you, 'Bob'. The difference is it actually applies to you.

>
> >>> BHCR is as worthless as your unwillingness to acknowledge you are wrong.
>
> >> I take it our buddy Dan here doesn't tune his buffer cache.
>
> > No. You should take it that he doesn't use worthless metrics to do so.
>
> Of course, some people think Buffer Cache Hit Ratio has nothing to do with
> Buffer Cache.

Now we get to the heart of your misunderstanding. No one has said the BCHR has nothing to do with the Buffer Cache. The issue is the reliability of using such a ratio as an indicator of database and buffer cache health and performance. There is none.

>
> >>> But given that your name is in valid, and your email is not valid, why
> >>> would anyone expect your opinion to be valid? Come on out of the closet.
>
> >> It appears valid names are more important to you than valid arguments.
>
> > Wrong again: You have neither.
>
> Again, you should not even pretend to be a judge.

He's not pretending. You, however ...

>
> > I wonder ... why would someone at San Diego State be so afraid to
> > acknowledge who they are? Perhaps some more research will answer
> > that question.
>
> Did someone just overdosed this guy? He is really lost in the dark now.- Hide quoted text -

If anyone is lost in the dark it's you, 'Bob', and you continue to prove that fact with posts such as this gem.

>
> - Show quoted text -

David Fitzjarrell Received on Sat Sep 08 2007 - 16:48:10 CDT