Wow....classic article!!! At least it delved slightly into more lower level. I just want to understand why "incremental checkpoint" is better than the normal complete checkpoint. I guess there is a balancing point. If u do too many small little things (like incremental does), u end up having less time to do the real work. Ideally zero incremental checkpoint is the best for performance - at the price of recovery. On the other hand, by understanding this "incremental checkpoint" better, I was looking for a more light-weighted version of it - "incremental incremental checkpoint". I will continue to understand the internals better. Thanks!!! On Wed, Mar 26, 2008 at 2:43 AM, Jared Still <jkstill@xxxxxxxxx> wrote: > In regards to checkpointing in general, ML note 147468.1 has a good > explanation and is easy to read as well. > > Note 438176.1 is somewhat outdated, but does have a high level explanation > of incremental checkpoints. > > However it does not answer such questions as "At what level is a incremental > checkpoint triggered?" > How old does a block have to be to be 'old'? > > > > > On Fri, Mar 21, 2008 at 6:26 PM, Peter Teoh <htmldeveloper@xxxxxxxxx> wrote: > > > > > Just attended a OU course in Performance Tuning. After hearing the > explanation on incremental checkpointing, I still cannot understand the > purpose/design of this incremental checkpointing. > > > > What is the key design, vs complete checkpointing, in Oracle database? > WHat is the tradeoff that it sacrificed, in order to achieve what it does? > > > > > -- > Jared Still > Certifiable Oracle DBA and Part Time Perl Evangelist > -- Regards, Peter Teoh -- //www.freelists.org/webpage/oracle-l