I would suggest trying to find the slon command on your machine using a command along the lines of:

find / -name slon

(this means: search from the root for a file named slon)

Good luck.

– You definitely do not need to drop/recreate the master database (or any database for that matter) to remove Slony. I always used used PgAdmin to do it (BTW, you can also use PgAdmin remotely, even if the particular box doesn’t have X for example – that’s how I used it). Looking through the documentation I found the DROP NODE command. I suppose that you would use it like this: create a shellscript with the contents shown below and run it (supposing that first_host is your master):

#!/bin/bash
slonik <<_EOF_
cluster name = slon_cluster_a;
node 1 admin conninfo = ''dbname=dbname host=first_host user=u password=p'';
node 2 admin conninfo = ''dbname=dbname host=second_host user=u password=p'';
drop node (id = 1);

Again, I never actually done this, so you should really try this in a test environment!

– Regarding starting the replication from scratch because the table structure changed: by regenerating the slonik script I mean re-running the SELECT I've shown, saving the result in a file (possibly the same file you used the first time, since that is of no use anymore) and running it. By restarting the slon daemon I just mean that you start the slon daemon on the slave again (you supposedly stopped it when you dropped the replications).

– Regarding failover: again, I never done it. I have been told that the config I generate needs to be augmented to provide two-way paths for this to work. I recommend that you try to look at the Slony documentation or ask on the Slony-I irc channel (#slony on irc.freenode.net). The later I found to be very helpful. Once I was helped by none other than the creator or Slony!