I'm interested in several questions wrt how people are implementing
Slony.  There may not be very many people doing this (yet)
but surprise me :-)

The questions are phrased for consultants implementing
slony for a client customer, but can also be answered for your own
implementation.  

Your feedback, however short, will help others tackle the problems
you have already resolved and maybe help you solve problems you
have not gotten to yet.

If you copy or send direct to me, I'll summarize the results back to the list.
If the summary is helpful we can put it up on the slony site and
I may also write an article based on the results. (Do not send
proprietry information :-)

1) I'm interested in how you are extracting the configuration
information from the client.  There is a lot of data to be gathered
before beginning regarding servers, connectivity, initial cluster
and failover preparations.  Do you talk?  Do you have a prepared
set of questions?  How else do you get this information?  Also, which
of these things are you deciding and which are your client deciding?

2) What tools are you using?  Custom slonik scripts? One or more of the tools
available on the slony site with or without modification to the tools and/or
the results?  Calling the Slony stored procedures directly from SQL?
Custom tools to generate slonik or stored procedure calls?   A good 
(or not) todo list?

3) Are you prepared to handle schema changes?  What does it mean to be
prepared for this?  What requirements do you have for the client wrt source
code control for schema changes?

4) What monitoring techniques are you using to ensure a) slony remains
running b) replication occurrs as expected?

5) Do you have pre-prepared failover plans?  As scripts? As a todo list?
As something else?

6) Are you replicating against different versions of Postgres?  How
do you handle Pg version upgrades?  How do you handle Slony upgrades?

7) How many nodes are in each cluster? How many levels is your cluster?
(Master -> Replica == 2 nodes, 2 levels) (Master -> Replica1 -> Replica2
3 nodes, 3 levels) (Master -> Replica1, Master -> Replica2 == 3 nodes,
2 levels)

8) What question should be asked here and wasn't? 

Thanks,

Elein
============================================================
elein at varlena.com        Varlena, LLC        www.varlena.com

          PostgreSQL Consulting, Support & Training   

PostgreSQL General Bits   http://www.varlena.com/GeneralBits/
=============================================================
I have always depended on the [QA] of strangers.