Re: Degrading performance - Mailing list pgsql-performance

From scott.marlowe
Subject Re: Degrading performance
Date
Msg-id Pine.LNX.4.33.0306021142180.12320-100000@css120.ihs.com
Whole thread Raw
In response to Re: Degrading performance  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-performance
On Mon, 2 Jun 2003, Tom Lane wrote:

> "scott.marlowe" <scott.marlowe@ihs.com> writes:
> > On Mon, 2 Jun 2003, Tom Lane wrote:
> >>> INFO:  Rel ifdata: Pages: 4887 --> 17; Tuple(s) moved: 776.
> >>> CPU 0.30s/0.35u sec elapsed 1.65 sec.
> >>
> >> That says you waited way too long to vacuum --- over two hundred update
> >> cycles, evidently.
>
> > Don't forget to crank up your fsm settings in $PGDATA/postgresql.conf as
> > well.
>
> The table's not very big though.  As long as he keeps after it with
> sufficiently-frequent vacuuming, it won't need much FSM space.

Yeah, but I got the feeling he was updating like 40 rows a second or
something.  Sufficiently frequent for him may well be constant. :-)


pgsql-performance by date:

Previous
From: Tom Lane
Date:
Subject: Re: Degrading performance
Next
From: Yusuf
Date:
Subject: Enabling and disabling run time configuration parameters.