Well, it's also a function of fast_start_mttr_target so reducing it
should lead to more frequent and short checkpoints. On the other hand,
it might stress IO subsystem more and give actually an opposite
result.
2006/5/24, Tanel Põder <tanel.poder.003_at_mail.ee>:
>
>
> Which version of Oracle?
>
> Oracle isn't very aggressive doing logswitch checkpoints anymore. You can
> easily test it out by having 3-4 redolog groups, doing logswitches and
> querying groups which have status = ACTIVE from v$log. The ACTIVE groups
> haven't checkpointed to disk yet. In 9i at least one log group was allowed
> to stay ACTIVE after log switch without initiating a checkpoint, in 10gR2
> all my 4 logfile groups were not checkpointed by default (had 1 CURRENT and
> 3 ACTIVE groups). But on next logswitch all three ACTIVE groups were
> checkpointed.
--
Best regards,
Alex Gorbachev
http://oracloid.blogspot.com
--
http://www.freelists.org/webpage/oracle-l
Received on Wed May 24 2006 - 06:50:22 CDT
