From: | Steve Crawford <scrawford(at)pinpointresearch(dot)com> |
---|---|
To: | "Matthew T(dot) O'Connor" <matthew(at)zeut(dot)net>, "Thomas F(dot)O'Connell" <tfo(at)sitening(dot)com> |
Cc: | Martijn van Oosterhout <kleptog(at)svana(dot)org>, pgsql-general(at)postgresql(dot)org |
Subject: | Re: Multicolumn Primary Key |
Date: | 2004-08-31 23:41:24 |
Message-ID: | 200408311641.24089.scrawford@pinpointresearch.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Tuesday 31 August 2004 3:49 pm, Matthew T. O'Connor wrote:
> Thomas F.O'Connell wrote:
> > On Aug 31, 2004, at 4:34 PM, Martijn van Oosterhout wrote:
> >> Otherwise, have you VACUUM ANALYZEd recently?
> >
> > Argh! That seems to have been it. I have pg_autovacuum running,
> > so it didn't occur to me.
>
> pg_autovacuum only knows about table activity that has taken place
> while it's running. Example: Table A has an analyze threshold of
> 1000, if you perform 999 updates on table A, then kill and restart
> pg_autovacuum, you will have to perform another 1000 updates before
> pg_autovacuum performs an ANALYZE. This is a large limitation of
> the current implementation.
>
> I tried to fix this for 8.0, but my autovacuum improvements didn't
> make the cut.
As a quick-'n'-dirty "fix", what about a command-line option to
autovacuum that tells it to vacuum analyze all tables on startup? At
least it would start in a somewhat known condition.
Cheers,
Steve
From | Date | Subject | |
---|---|---|---|
Next Message | Matthew T. O'Connor | 2004-08-31 23:44:33 | Re: Multicolumn Primary Key |
Previous Message | Steve Crawford | 2004-08-31 23:36:55 | Re: expected both swapped tables to have TOAST tables |