Hello, all,

Slony 1.0.5, Pg 7.4.6, Linux x86.  Growing load is forcing us to add a 
new machine into our Slony master->slave configuration, and before I 
start firing slonik commands off, I'd like to make sure that I 
understand exactly what I will be doing.

Assume that the master machine is node id #1, and the existing client, 
id #2; in addition, master provides set #1, to which the client 
subscribes.  We will add the new machine in at node id #3.  Also, the 
client nodes will not have any contact with each other -- our failover, 
for various reasons not related to Slony, is un-automatable (sic.)

1. Copy the existing schema (sans the _${CLUSTER_NAME} namespace) to 
the new machine;

2. run the following script on the master;

NODE 3 ADMIN CONNINFO = '${SLAVE2_CONNINFO}';

STORE NODE (id = 3, comment = 'New slave node' );
STORE PATH (server = 1, client = 3, conninfo = '${SLAVE2_CONNINFO}';
STORE PATH (server = 3, client = 1, conninfo = '${MASTER_CONNINFO}';
STORE LISTEN (origin = 1, receiver = 3);
SUBSCRIBE SET (id = 1, provider = 1, receiver = 3, forward = no);

3. start the slon daemon on the client machine;
4. and stand back.

Is this complete?  Are there any caveats?

jfb

PS: I would be interested in contributing some of our experiences in 
migrating from dbmirror to Slony to a documentation project; does such 
a creature exist?

-- 
James Felix Black
Programmer, iParadigms LLC
(510) 287-9720 x 250