Re: Oracle 10g RAC performance

On Apr 27, 11:46 am, Valentin Minzatu <valentinminz..._at_yahoo.com> wrote:
> On Apr 27, 8:29 am, Robert Jaroszuk <z..._at_iq.no.spam.please.pl> wrote:
>
>
>
>
>
> > Hello.
>
> > How (approximately) performance increase can I expect when i migrate one
> > dual-core Xeon server into two dual-core Xeon servers in RAC ?
> > Do you know any benchmarks comparing vertical-scaling vs
> > horizontal-scaling of Oracle servers?
>
> > What about two-server RAC vs three-servers RAC ?
> > How performance increase can I expect ? 30% ?
>
> > Best regards,
>
> > --
> > ... Robert Jaroszuk ...
> > GCS/IT/O d? s: a- C++ ULB++++$ P+ L++++$ E- W++ K- N+ DI+ V-
> > w M- PS+ PE Y(+) PGP-(+++) t-- 5? X R !tv b++>++++ D- y+ G++
> > .http://zim.iq.pl/. RJ735-RIPE .http://zim.iq.pl/photo/.
> > .. The superior warrior wins without fighting -- Sun Tzu. ..
>
> > -> New photos:http://zim.iq.pl/photo/
>
> My two cents: if your application scales very well in a single node
> configuration, then you should expect it to scale in a cluster
> configuration as well (up to 90-95% according to Oracle). If your
> application does NOT scale well in a single node configuration, then
> forget about RAC: it will only make things worse. Also keep in mind
> that the most penalty you pay in a RAC environment is when you add the
> 3rd node. An interesting presentation can be found on RAC SIG website:http://www.oracleracsig.org/pls/htmldb/Z?p_url=RAC_SIG.download_my_fi....
>
> Cheers,
> Valentin

Well geez let's not start putting out some half way reasonable interpretations around cdos it's going to spoil the fun of the oracle marketing fossils.

There's a reasonable smart guy who wrote "Why you probably don't need RAC" ... there's somebody who is clued in to the real world.

The 90 to 95 percent thing sounds more like a "forget about it" with carefully pre-arranged benchmarking of painfully created non quite real world applications. Received on Fri Apr 27 2007 - 19:48:53 CDT