On Sat, 2009-05-30 at 00:08 -0400, Andrew Sullivan wrote:
> On Fri, May 29, 2009 at 03:37:04PM -0700, Melvin Davidson wrote:
> > >>  sequence ID 1 has already been assigned
> > 
> > There's your big clue! You probably have two sequences somewhere in
> > the config with the same sequence ID. 
> 
> I agree this has to be the problem.
> 
> > sequences with the same ID as a table, I always start my sequences
> > with an offset of 1000 from table ID's just to make tracking easier.
> 
> Gee, and then what do you do when you end up with 1000 tables, because
> you have to do with table rotation due to huge success of your
> incredible application?  Integers in int32 space are pretty much
> free.  If you're trying to save yourself trouble this way, offset by
> more than 1000 -- 10 000 or even 100 000 isn't a bad idea.
> 
> Note I'm not objecting on principled grounds that "you might need 
> 10 000 tables" or anything.  This is just a prudential remark that an
> offset of 1 000 is a little too close, in my experience.

I wonder though if there will be a limiting factor in Slony itself
though.  I recall that node numbers had some sort of artificial ceiling
around 9000 or so, due to how they are used internally. 

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