On Mon, 2009-11-23 at 11:25 -0500, Christopher Browne wrote:
> Laurent Laborde <kerdezixe at gmail.com> writes:
> > i had a slave that was lagging by 2~5mn.. slowly but surely, without
> > high load coming from the read request.
> > overall... the average load was good. (around ~4 on an octocore).
> >
> > After many try i found that : vacuum analyze pg_listener; instantly
> > solved the problem.
> > (recovering from a 2mn lag in just a few second, and now synched with ~3s lag).
> >
> > Any tought about why i could have a bloated pg_listener ?
> >
> > postgresql 8.3.6
> > Slony 1.2.15
> > pgbouncer
> 
> Well, Slony-I does invalidate quite a lot of pg_listener tuples, so it's
> not a surprise for this to happen if pg_listener isn't getting vacuumed.
> 
> The cleanup thread is supposed to vacuum pg_listener once in a while
> (I'd expect about every 10 minutes, if slon runs with defaults), so I'd
> wonder why that, or some other vacuum, is not happening.

If the vacuums are running, do you have a lot of long running
transactions against this database?

-- 
Brad Nicholson  416-673-4106
Database Administrator, Afilias Canada Corp.