Re: Log Buffer in SGA <---> Online redo files

In article <s3jrjenv3ef23_at_corp.supernews.com>,   "Jonathan E Miller" <jmiller_at_fnol.net> wrote:
> I'm about to create a new database (my first real production one)
and was
> wondering whether there are recommended sizing relationships between
the log
> buffer in SGA and the redo files on disk. I've set the redo files to
5M and
> there are 3 mirrored groups of them and I've set the log buffer to
500K.
> Will this work? ANd I mean will it work efficiently? The DBA guides
dont
> have much in the way of log buffer sizing recommendations. Thanks very
> much -
>
> Jonathan
>

I think your log buffer setting is probably large enough. Indead I have recently listened to one expert who will argue that it is too large since you will be wasting most of the allocated space.

You log buffer will be fushed on commit, when 1/3 full, and due to checkpoints. If your application is an OLTP with a lot of small transactions then the rdbms will be flushing the buffer constantly so you may end up not being able to use all of the allocated space.

I have not looked at the log buffer statistics in a while so I can not remember off the top of my head which ones to look at but the statistics are in v$sysstat with a name like 'redo%'.

--
Mark D. Powell -- The only advice that counts is the advice that  you follow so follow your own advice --

Sent via Deja.com http://www.deja.com/
Before you buy. Received on Tue Nov 23 1999 - 08:23:09 CST