Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.misc -> Re: general tuning results

Re: general tuning results

From: Daniel A. Morgan <damorgan_at_exesolutions.com>
Date: Mon, 21 Jan 2002 11:52:56 +0000
Message-ID: <3C4C0118.2E78BC24@exesolutions.com>


Actually many more than that. But be sure to include network traffic and speed and hard disk I/O capacity.

But one thing that will really limit performance is RAID5. Oracle does not recommend it. And the redo logs should absolutely NOT be RAIDed.

Daniel Morgan

Ed prochak wrote:

> "Frederic Mockel" <frederic.mockel_at_gillardon.de> wrote in message news:<3c4c158d$0$227$4d4ebb8e_at_businessnews.de.uu.net>...
> > Hi newsgroup,
> >
> > i have a little problem,
> >
> > we have a machine on w2k with 2 processors and 2GB ram and hardisk at RAID5.
> > I tuned a database for this machine.
> >
> > Know i think on the question if the tuning results portable for example a
> > Sun Solaris 2 cpus, 2 GB of ram and a hardisk at RAID 5
> > or generally asked
> >
> > If the big parameters like CPUs Ram and HDD similar the same and only the OS
> > is different, can i transport the tuning results to this maschine ?
> >
> > Waiting for your response
> >
> > Freddy
>
> That would be a good place to start. Just be aware that even with
> identical hardware you have at least two variables different:
> OS/ORACLE version and DATA.
>
> The designs of those OS's WIN2K and SOLARIS are very different so
> there will be some effect on the operation of ORACLE. And even ORACLE
> had to be built on each platform possibly with different tools with
> different optimization features.
>
> How the data was installed also differs. One one machine the data may
> have been loaded bit by bit over a long period of time in a production
> environment. The you go to test on another machine, so you export and
> import the data which can lay the daya out on the disc drives very
> differently.
>
> The bottom line is to start with what you have for tuning parameters,
> but then continue to experiment. Tuning seems to always require a
> great deal of testing and experiment.
>
> Good luck.
>
> Ed Prochak
Received on Mon Jan 21 2002 - 05:52:56 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US