I need a free (as in freedom) multimaster replication solution for PostgreSQL.
As I know no such thing exists at this moment, I think a 'star replication
topology' can do the work for me (I will explain it better in my next mail to
the list). But I need to know some details about the Slony-I behaviour when a
update sql query that do not really alter the data is performed.

As my English is not quite good, I'll write two example cases to see if someone
can tell me what Slony-I would do in each one:

My example table have 2 fields ('key' that is pk and 'name') and just one record
Initial state (master and slave databases both have this data):

Example_table
-------------
key | name
-------------
 4  | juan

1) In the master node this query is performed:
  UPDATE Example_table SET name = 'juan' WHERE key = '4'

  It's the same query performed in the slave node?
  (Note that, in fact, it's not necessary)

2) Slony-I is stopped in the slave host. Then (without replication) the
  following query is performed in the slave database:
  UPDATE Example_table SET name = 'pepe' WHERE key = '4'

  Then, Slony-I is restarted and the above query is performed in the master node
  (in fact, before the sync take place the data is now identical in both
  databases, but the master does not know it).

  Will be a new update statement executed in the slave database?
  (Note that, in fact, it's not necessary)

I have the same questions about insert statements but I don't want to write a
kilometric email.

Greetings

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.