Re: Redo Log Buffer Contention

In Tuning Oracle, the stress is on ensuring that there are no wait situations with respect to the redo log buffer cache. To monitor this, you look at v$latch where the latch name is like '%redo%'. The sum of immediate_misses should be less than 1% of immediate_gets or the sum of misses should be less than 1% of gets. Anything greater suggests contention for the redo allocation or the redo copy latch.

If you are interested in some scripts to query this information, let me know.

Glenn Stauffer
DBA
Swarthmore College

In article <49hics$mj_at_news.nstn.ca>, nstn3240_at_fox.nstn.ca wrote:

>The documentation I have states that the value of 'redo log space
>requests' (from the v$sysstat), should be near 0. This is
>accomplished by adjusting the LOG_BUFFER parameter in the init.ora
>file.
>
>The question is, what's 'near 0'? I'm about 50 is that considered
>near zero, what's the scale?
>Gerry Flemming
>nstn3240_at_fox.nstn.ca
>Datacor/ISM
>Moncton, NB CANADA