--- On Wed, 7/4/10, Glyn Astill <glynastill at yahoo.co.uk> wrote:

> From: Glyn Astill <glynastill at yahoo.co.uk>
> Subject: [Slony1-general] Bloated sl_seqlog_idx index
> To: slony1-general at lists.slony.info
> Date: Wednesday, 7 April, 2010, 12:04
> Hi Guys,
> 
> Slony 1.2.20 / Postgres 8.4.3
> 
> I'm seeing a fair bit of index bloat on the sl_seqlog_idx,
> bizzarely only on the one server in a 5 server setup. 
> The server involved is a subscriber to the origin and a
> provider to one downstream subscriber.
> 
> The server in question was recently moved from our local
> network (where I'd not seen this happen over 3 months) to a
> hosting centre on the other end of a vpn, which I've noticed
> drops occasionally causing slony connections to drop out and
> have to reconnect. Could this be a problem?
> 
> Does anyone have any idea why this might be or why only the
> one server is affected?
> 
> I've always disabled autovacuum against slony tables as I
> thought slony did it's own vacuuming, is this still the
> correct thing to be doing or should I enable autovacuum?
> 
> Glyn
> 
> 

I figured it out.  It appears I was being a total plank, ignorant to the fact pg_autovacuum was switched for an autovacuum_enabled parameter on the table in 8.4.