> -----Message d'origine-----
> De : Alan Hodgson [mailto:ahodgson at simkin.ca] 
> Envoy? : mercredi 3 novembre 2004 17:47
> ? : slony1-general at gborg.postgresql.org
> Objet : Re: [Slony1-general] One cluster for all 
> replications... or not?
> 
> On Wed, Nov 03, 2004 at 05:41:48PM +0100, J?r?mie LOPEZ wrote:
> > What is the best perspective concerning several 
> replications running 
> > on shared machines?
> > 
> > Let's consider I have - for the same project - 2 PostgreSQL servers 
> > (on 2 distinct machines): I want the server 1 to be the master for 
> > database A and the slave for database B, and the server 2 to be the 
> > master for database B and the slave for database A.
> > 
> > What does follow the "slony way" the most between :
> > - defining two clusters (one for each database to replicate) with 2 
> > nodes each,
> > - defining a unique clusters with 4 nodes (and making the correct 
> > subscriptions)?
> >
> 
> I think you have to setup 2 clusters (and end up running 4 
> slon's) ,

Well, in fact I have 4 slons in each case, since I have 4 nodes (i.e. 4
databases), from the same cluster or not...

> because the cluster schema itself goes into the 
> database where the replicated data comes from.

(and in the database where the replicated data goes too...)

> You can't 
> talk to 2 databases from the same connection.

I don't understand the constraint you're talking about... I don't see the
differences having 1 or 2 clusters in terms of connections...

> 
> --
> How many of you believe in telekinesis? Raise my hand...
> 
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at gborg.postgresql.org
> http://gborg.postgresql.org/mailman/listinfo/slony1-general
> 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gborg.postgresql.org/pipermail/slony1-general/attachments/20041103/326bc26e/attachment.html