Re: redo latch contention

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Wed, 22 Sep 1999 08:25:57 +0100
Message-ID: <937985400.24968.0.nnrp-14.9e984b29@news.demon.co.uk>

log_simultaneous_copies has become a
hidden parameter. I suspect is now defaults to 2 x CPUs, even on a single CPU system.

Also, the write algorithm seems to have changed, if you look at v$sysstat, 'redo small copies' has disappeared - hence the redundancy of the log_small_entry_max_size.

Jonathan Lewis
Yet another Oracle-related web site: http://www.jlcomp.demon.co.uk

Jonathan Gennick wrote in message
<37ed1e58.10878426_at_netnews.worldnet.att.net>...
>I was doing some research today and noticed that the 8i
>release of Oracle has obsoleted the LOG_SIMULTANEOUS_COPIES
>and LOG_SMALL_ENTRY_MAX_SIZE parameters. The tuning manual
>however, says to adjust these in the event that you are
>experiencing redo latch contention. With these parameters
>gone, is there anything left that you can do about this type
>of latch contention? Does Oracle consider it to be an
>insignificant problem?
Received on Wed Sep 22 1999 - 02:25:57 CDT