On Wed, Apr 23, 2008 at 9:27 AM, John Gardner <john.gardner@tagish.co.uk> wrote:
> We have two PostgreSQL servers (8.2) running in a cluster.
>
> We have autovacuum switched on on both servers and also we are running the
> following as a cron job;
>
> Server 1:
> 30 0,2,4,6,8,10,12,14,16,18,20,22 * * * /usr/bin/vacuumdb --all --analyze
You can change that to
30 0-59/2 * * * ...
and
30 1-59/2 * * *
(I think the second one does the odd minutes. might need to test it.
> 1) Is using the autovacuum daemon and running vacuumdb from a cron job
> overkill?
Maybe. It's probably better let autovacuum handle most of the db, and
use a single vacuum analyze on the one really busy table.
It's possible that you're using up too much of your I/O bandwidth
cleaning a bunch of table that don't need it. As Joshua mentioned,
it's probably a good idea to look at 8.3 due to its ability to run >1
autovacuum thread at a time.