Simon Riggs wrote:
> All I'm saying is *if* we put scheduling inside Postgres for autovacuum
> *then* we should make it general purpose scheduling.
>
> If anybody uses the argument that "we have external schedulers, so don't
> put them in the database" then that argument applies equally to
> scheduling autovacuum. It's easy to turn autovacuum on/off via an
> external scheduler, yet look upthread and see how many people think it
> should be in the database.
>
> Whichever way you think the decision should go, the same arguments apply
> to scheduling autovacuum and scheduling other database maintenance
> tasks.
>
>
OK, I agree with that.
cheers
andrew