Big Checkpoints

From: kirsty-purcell <kirsty-purcell_at_supanet.com>
Date: Sun, 8 Apr 2001 09:36:29 +0100
Message-ID: <0v7pa9.npb.ln@meszaros.user.cis.dfn.de>

Hi,
I've got a small (2Gb), database on 8.1.6 running on NT4. The server has 3CPU, 2.5Gb mem and RAID1 for reo, RAID1 for archlogs and RAID5 for data.

The problem I'm getting is that a log switch checkpoint takes 30 mins to complete, even though it doesn't look as though its doing anything.

I've tried setting log_checkPoint_timeout but that doesn't seem to do anything (how does this work).
Is there any recommended setting for db_block_max_dirty?

The only disk activity appears to be a write to the control file every few seconds. What does Oracle do in a checkpoint?

Any help would be great. Received on Sun Apr 08 2001 - 03:36:29 CDT