TO be truthful, I did not check that.
I arrived at the conclusion because although I have 10 autovacuum
threads, none of the tables had been autoanalyzed or autovacuumed in
hours (and due to the nature of the activity they do so every minute).
If I ever see this behavior now I know what to check for :) Thx.
> -----Original Message-----
> From: Kevin Grittner [mailto:Kevin.Grittner@wicourts.gov]
> Sent: Saturday, May 01, 2010 8:27 AM
> To: Benjamin Krajmalnik; pgsql-admin@postgresql.org
> Subject: Re: [ADMIN] Autovacuum stopped running
>
> "Benjamin Krajmalnik" <kraj@illumen.com> wrote:
> > I had a very strange occurrence 2 days ago where autovacuum
> > appears to have stopped running.
>
> What did you see which led you to that conclusion? Did ps show the
> 'autovacuum launcher process'?
>
> > I did not see any error messages in pg_log. The problem caused a
> > file which is usually around 50Mb but which gets updated
> > extensively to grow to 105GB, which in turn brought the server's
> > performance down to unusable levels.
>
> Were there any long-running transactions showing in
> pg_stat_activity? (That would include transactions showing '<IDLE>
> in transaction'.)
>
> -Kevin
