Re: Checkpoint not complete

On Mon, 4 Jan 1999 17:02:45 -0000, "Pedro Tavares" <pedrotavares_at_cmvm.pt> wrote:

>My database has lots of activity, and if I monitor de alert log, I'm
>constantly receiving the message - checkpoint not complete. I believe that
>this is affecting the overall performance. What can I do to avoid this? I
>have 4 redo log files of 1M each.
>Thanks in advance.
>
>Pedro Tavares
>Lisbon, Portugal
>

Increase the size of the online redo logs, or alternatively add more logs. Online redo logs in oracle are used in a circular fashion. What's happening is that each time you do a log switch, a checkpoint is initiated. Your database is trying to cycle back to an online redo log before the checkpoint associated with that log switch has completed. By increasing the online redo log size your're giving it more time to complete.

You say your logs are 1m x 4. 1m is way too small. Try it at about 20m and see if that eliminates the problem. If not, start increasing by 10m increments and stop when the problem disappears.

On one of my OLTP system, we're running 100m x 3. The reason our logs are so big is to reduce checkpointing. There's a lot of I/O associated with a checkpoint.

Also, make sure you've set your init parameters log_checkpoint_interval, and log_checkpoint_timeout correctly. I always set the timeout to 0 and interval high enough so that the only checkpoints I get are when I do a log switch.

        Chuck
--
Chuck Hamilton
chuckh_at_safeaccess.net

Never share a foxhole with anyone braver than yourself! Received on Mon Jan 04 1999 - 12:56:50 CST