Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: Another Oracle "Myth"?
"Niall Litchfield" <n-litchfield_at_audit-commission.gov.uk> wrote:
> "Noel" <tbal_at_go2.pl> wrote in message
> news:ec30e927.0311210141.3369e057_at_posting.google.com...
> > "Anurag Varma" <avdbi_at_hotmail.com> wrote in message
> news:<yDfvb.529$Sm1.65_at_news02.roc.ny>...
> > > "Geomancer" <pharfromhome_at_hotmail.com> wrote in message
> news:cf90fb89.0311201853.126b1516_at_posting.google.com...
> > >
> > > He does not say that 99.9% hit ratio is always bad (which you seem to
> > > be
> interpreting).
> > > However he does seem to be saying one should NOT rely on 99.9% hit
> > > ratio
> to make the judgment that
> > > the database performance is good. The fact might just be the
> > > opposite.
> >
> > Not hard to imagine bad sql query slowing the database performance.
> > Hit ratio hides number of memory/disk reads.
> > If you would load all datafiles into memory and database buffers the
> > hit ratio would always be 100%.
>
Do dirty blocks really generally get reread after being written to disk? If it's written out because someone needs a free cache block, then sure it needs to be re-read next time it's needed. But if it gets written out due to a checkpoint or just a bored DBWR, isn't it still available in memory, just marked clean rather than dirty?
Xho
(If you saw the first post, oops. DBWR, not LGWR)
-- -------------------- http://NewsReader.Com/ -------------------- Usenet Newsgroup Service New Rate! $9.95/Month 50GBReceived on Fri Nov 21 2003 - 13:27:10 CST