On Thursday, November 01, 2012 05:40:23 PM Alban Hertroys wrote:
> On 1 November 2012 17:19, Shaun Thomas <sthomas@optionshouse.com> wrote:
> > On 11/01/2012 10:28 AM, Kevin Grittner wrote:
> > Based on my past experience with 8.2, and my understanding of 9.1, I
> > moved autovacuum_freeze_max_age up to 650M so we'd never get a mid-day
> > freeze. And the default for vacuum_freeze_table_age is 150M, which I
> > hadn't changed.
>
> Instead of attempting to postpone freeze until beyond the life
> expectancy of our universe, what you probably should have done is
> vacuum more often so that vacuum has less work to do.
Thats not really possible with freeze vacuums. When the table is older than
the applicable freeze age its scanned completely instead of only the parts
that are sensible according to the vacuum map. The more expensive scans really
only happen when they make sense...
Andres
--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services