Re: Redo latch question

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Fri, 9 Jul 1999 20:34:53 +0100
Message-ID: <931549206.7910.0.nnrp-08.9e984b29@news.demon.co.uk>

Also,

If you have multiple redo copy latches (which is almost invariably the case), a process may try to grab a latch, miss because it is in use, grab the next one, miss etc.

You may find that willing_to_wait gets on redo latch allocation is quite close to 'immediate gets' MINUS 'immediate misses' on redo copy latch, plus a fudge factor close to the 'redo small copies' sysstat statistic.

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

akolk - gelrevision.nl wrote in message <3786447C.5164A290_at_gelrevision.nl>...
>LGWR also tries to get the copy latches before it writes out the redo.
>
>webbo wrote:
>
>> If every redo copy must obtain an allocation latch, but does not
necessarily
>> need a copy latch (depending on settings of log_small_entry_max_size &
>> log_simultaneous_copies parameters), how come my gets+im_gets are greater
for
>> the redo copy latch???
Received on Fri Jul 09 1999 - 14:34:53 CDT