Re: Question about sizing log files

"Sanjay Mishra" <oraskm_at_yahoo.com> wrote in message news:eca56b4a.0304022000.58fae4ad_at_posting.google.com...
> The proper sizing of redolog files is a trade-off between better
> runtime performance and better recovery performance.
>
> Bigger log files --> less frequent log switches --> better runtime
> performance.
>
> Bigger log switches --> less frequent checkpoints(provided you have
> high log_checkpoint_interval) --> more time required during instance
> recovery.
>
> Smaller log files --> frequent log switches --> poorer runtime
> performance.
>
> Smaller log files --> frequent checkpoints --> less time required for
> instance recovery.
>
> ** a checkpoint occurs at every logswitch.
>

Hi Sanjay,

The above was true several years ago but now that Oracle uses incremental checkpointing as it's default mechanism, the issue of redo log size and checkpoint performance vs instance recovery times is somewhat redundant.

To Chuck, based on the significant amount of redo you generate, I would recommend increasing the size of your redo logs to be around the 500M mark Then tune your fast_start_mttr_target or fast_start_io_target to get the desired balance between instance recovery times and checkpoint overheads. This can be easily determined by checking out your statspack output during a busy period to see what buffer writes due to checkpoints have occurred and what are the estimated instance recovery times.

Cheers

Richard Received on Thu Apr 03 2003 - 05:30:32 CST