Re: Checkpoint time

On Sun, 4 Jan 2004 14:19:38 +0100, "EP" <ep_at_NOSPAMplusnet.pl> wrote:

>Hello @,
>
>I have encountered the following problem: during an import on a test server
>some of the checkpoints take up to 50 minutes to complete - they are
>triggered by log switches - the size of redo logs is 500 MB.
>When I start the import everything seems to be OK, but afted a considerable
>amount of the import time (i.e. 24-36 hours) this problem occurs.
>I have browsed through the metalink but found no bug/solution.
>The platform: HP-UX 64bit, Oracle: 8.1.7.4
>
>Thanks for any clues,
>
>Best Regards,
>EP
>

Try enabling incremental checkpointing, by changing log_checkpoint_interval to something meaningful. Also (you don't provide that much information do you ) try to determine whether you have any disk bottleneck (eg multiple copies of a redolog on one single disk)

--
Sybrand Bakker, Senior Oracle DBA