In response to Devrim GÜNDÜZ <devrim at gunduz.org>:
> 
> What is the trick to setup cascading slaves with Slony-I? I'm probably
> blind, since I did not see anything in the document. What I want is:
> 
> pg1 will replicate to pg2 and pg2 will replicate to pg3.
>  
> I setup two different slon_tools.conf for these purposes, initialized
> slony clusters, started daemons for each of them, and subscribed sets.
> 
> I could perform initial sync -- however, when I insert a test data from
> pg1, I can see it pg2 -- but not in pg3. 
> 
> What am I doing wrong?

Not providing detailed enough information for anyone to help?

Make sure you have a slon process starting for each of the nodes (usually,
they run _on_ the node, but that's not required).  I.e, you should have
3 slon.conf files and the scripts starting slons up.

2x check what's in [cluster_name].sl_path, [cluster_name].sl_subscribe, and
[cluster_name].sl_status ... those tables may give you hints as to what's
not right.

Check the contents of your slon logs.  They can be pretty cryptic, but
sometimes the output will outright tell you what's wrong.  (Nothing like
a nice "no pg_hba.conf entry for user slony" message to clue one in)

I have this working in 11 different (separate) configurations, so I know
it works and is reliable ;)

-- 
Bill Moran
http://www.potentialtech.com
http://people.collaborativefusion.com/~wmoran/