Re: Optimization question

From: Anton Dischner <dischner_at_med.uni-meunchen.de>
Date: Tue, 23 Aug 1994 12:41:35 +0200
Message-ID: <dischner-2308941241350001_at_gkc12a.klch.med.uni-muenchen.de>


In article <33bhqr$8v1_at_cnj.digex.net>, pkuba_at_cnj.digex.net (Pete Kuba) wrote:

I think the greatest impact on performance is main memory (especially for your case).

Try to make the SGA greater
(its the db_block_buffers parameter in your init.ora).

Don't make it to big, if it gets greater than main memory you will see heavy threashing and a dramatic loss of performance.

Hope this helps,

Toni

> The company that I'm working for has Oracle 7.0.13 running on SCO Unix on
> a 486/66 w/ 32Mb. If there are more than a few users logged on,
> performance starts to really lag. Since I'm a programmer and not a dba
> (and since the company doesn't want to pay a dba), I'm wondering how to
> answer my boss' question, "what do we need make things fast?"
>
> These are a few things I've thought of:
>
> (1) Tune Oracle. Right now everything's set to the default
> values.
>
> (2) Get some more hard drives and split tablespaces across
> them. Right now everything's off of one drive, so
> I expect that contention is aweful.
>
> (3) Get a faster CPU, like a Pentium.
>
> I tuned most of the sql, and created indexes where they were helpful.
> The logic's close to as fast as it's gonna get.
>
> Also, I'd like to know the expected gains of the options. If (3) will
> increase performance by 100%, and (1) will only yield 20%, I'll go with
> (3).
>
> Thanks for all opinions,
>
> -Pete

+--------------------------------------------------------------------+
| Anton Dischner, DBA and system-programmer    Phone: +49 89 70953202|
| Institut fuer Klinische Chemie               Fax  : +49 89 70958888|
| Klinikum Grosshadern                         Home : +49 89 6254060 |
| Ludwig Maximilians Universitaet Muenchen     Handy: +49 172 8388880|
| 81366 Muenchen                                                     |
| Marchioninistr. 15                                                 |
| Germany                          Mail: dischner_at_med.uni-muenchen.de|
+--------------------------------------------------------------------+
Received on Tue Aug 23 1994 - 12:41:35 CEST

Original text of this message