Re: 10g Automatic Checkpoint Tuning

Well, checkpointing and log switches may or may not be related. A log switch will cause a checkpoint, but you can have ongoing checkpoints without log switching.

You don't say how often your logs are switching.

RF

• Raj Mareddi <yoursraju007_at_gmail.com> wrote:

> Guys,
>
> The following are the top-4 waits after CPU Time in
> my AWR reports
> during a batch window.
>
> log file switch (checkpoint incomplete)
> free buffer waits
> log file switch (private strand flush incomplete)
> log file parallel write
>
> Just realized that log switches are happening very
> frequent.. each log
> member is 300M and 3 groups. this size seems to be
> okay.. other
> database has 100M member and the same batch job
> runs good there.
> First wait says, checkpoint incomplete.. I have
> started looking at
> checkpoint tuning.. I have uncommented the following
> parameters
>
> - LOG_CHECKPOINT_INTERVAL
> - LOG_CHECKPOINT_TIMEOUT
>
> and set FAST_START_MTTR_TARGET to 60... Still dint
> see any big
> difference. Has anyone started using
> FAST_START_MTTR_TARGET for
> automatic checkpoint tuning ? Document says if we
> just need checkpoint
> tuning, we need to set this higher.. but it doesn't
> say how higher..
> Im thinking of going back to old boat of setting
> using interval value
> or FAST_START_MTTR_TARGET to a very high no#. . Any
> comments please.
>
> Automatic Checkpoint Tuning in 10g: 265831.1
> REDO LOGS SIZING ADVISORY: 274264.1
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

Robert G. Freeman
Author:
Portable DBA: Oracle (Oracle Press)
Oracle Database 10g New Features (Oracle Press) Oracle9i RMAN Backup and Recovery (Oracle Press) Oracle9i New Features (Oracle Press)
Oracle Replication (Rampant Tech Press)
Mastering Oracle8i (Sybex)
Oracle8 to 8i Upgrade Exam Cram (Coriolis <RIP>) Oracle 7.3 to 8 Upgrade Exam Cram (Coriolis <RIP>)

--
http://www.freelists.org/webpage/oracle-l