Re: Redo log buffer size

joel garry wrote:
> Niall Litchfield wrote:
> > Charles Hooper wrote:
> > > oracle dba wrote:
> > >> Hi all,
> > >>
> > >> According to books redo will to written to file after 1MB is full. So
> > >> is there any way that our redo log buffer is equal to or more than
> > >> 10MB.
> > >> If yes then why ..........................
> > >> if u know any metalink addr pls inform
> > >
> > > Granule size can force the LOG_BUFFER value to 16MB, even if LOG_BUFFER
> > > is explicitly set to 1MB in the initialization parameters.
> > >
> >
> > I think you'll find that only dynamically resizable components of the
> > sga are allocated in granules and that LOG_BUFFER does not meet that
> > requirement.
>
> Also see metalink Note:373018.1 and not-a-bug 4930608
>
> Learn something new every day.
>
> jg
> --
> @home.com is bogus.
> Those kookie banks: http://catless.ncl.ac.uk/Risks/24.46.html#subj7

Metalink note 351857.1 seems to explain what I posted (second posting any way) very well - the explanation is a bit different than I was expected, but seems to be consistent with my 4 database test in a previous post. Niall, you are correct - log_buffer is not allocated in granule size, but almost granule size would be a true statement according to Metalink.

Charles Hooper
PC Support Specialist
K&M Machine-Fabricating, Inc. Received on Mon Nov 06 2006 - 18:56:56 CST