On 22 Aug 2003 at 12:35, Claudio Lapidus wrote:
> Bruno Wolff III wote:
> > On Fri, Aug 22, 2003 at 12:17:41 +0530,
> > Shridhar Daithankar <shridhar_daithankar@persistent.co.in> wrote:
> > >
> > > Idea of autovacuum is to reduce load on vacuum full. If you set
> shared_buffers
> > > higher and FSM properly for he update/delete load, autovacuum is
> expected to
> > > catch most of the dead tuples in shared cache only. If it is successful
> in
> > > doubling the frequency on vacuum full, that's a big win, isn't it?
> >
> > If you run a normal vacuum often enough, you shouldn't need to regularly
> > run vacuum full.
>
> Hmm, here we have a certain table, sort of FIFO, rows get inserted all the
> time, lay there for a couple of hours and get deleted "the other end
> around". We run normal vacuum almost constantly, but the table keeps
> growing. We had to implement a 'vacuum full' once a week to keep it under
> control.
I think you could benefit from some kind of emulated partitioning.. If there
are large number of rows getting deleted at the end of two hours, you could
just drop that sub table..
Bye
Shridhar
--
We come to bury DOS, not to praise it.(Paul Vojta, vojta@math.berkeley.edu,
paraphrasing a quote of Shakespeare)