Re: Oracle and Memory

From: Joel Garry <joelga_at_rossinc.com>
Date: 1996/08/16
Message-ID: <1996Aug16.234226.19577_at_rossinc.com>#1/1


In article <4utgiv$o6a_at_nntp.interaccess.com> andre_at_interaccess.com (Andre T. Allen) writes:
>In article <4usk08$p7s_at_newsserv.cs.sunysb.edu>, trubisz_at_cs.sunysb.edu (Joe Trubisz) wrote:
>>This might sound like a totally idiotic question, since it
>>flys in the face of reason. But, an associate of mine just
>>came back from a Oracle class where the statement was made that
>>'having too much memory can cause problems with Oracle'. I
>>thought that this was totally absurd and the person in the
>>class didn't ask the obvious question, why?
>>
>>Does this make sense to anyone or was the instructor out to lunch?
>>
>>Thanks,
>>Joe
>>
>
>Too Much memory? In the day and age where it's slowly becoming chic to cache
>a whole database in memory, "too much" memory seems somewhat absurd.
>
>I have however, heard of instances where too much memory was allocated to the
>Oracle shared pool. This may result in some performance degradation if the
>shared pool is so big that it would actually take longer to search the shared
>pool for a cached SQL statement than it would be for the statement to be
>re-parsed and executed.

And if the SGA is so big it causes swapping, while a big part of the SGA isn't even being used, that is certainly too much memory. That will cause problems.

select sum(gets), sum(getmisses) from V$ROWCACHE. if sum(getmisses)/sum(gets) << .1, you have allocated too much memory to the data dictionary and library caches. Probably won't cause problems, but a waste of an expensive resource better allocated to users.

>
>Is this person going to Oracle Education, or to a 3rd party? I have heard
>some stories about certain instructors not knowing their a$$ from a hole in
>the ground.
>
>Regards -
>
>***************************************************************
>** Andre T. Allen * andre_at_interaccess.com **
>** SysAdmin/DBA * 74262.3572_at_compuserve.com **
>** Reid Psychological Systems * 312.938.9200 (Voice) **
>** Chicago, IL * 312.294.0104 (Fax) **
>***************************************************************

-- 
Joel Garry               joelga_at_rossinc.com               Compuserve 70661,1534
These are my opinions, not necessarily those of Ross Systems, Inc.   <> <>
%DCL-W-SOFTONEDGEDONTPUSH, Software On Edge - Don't Push.            \ V /
panic: ifree: freeing free inodes...                                   O
Received on Fri Aug 16 1996 - 00:00:00 CEST

Original text of this message