This tutorial assumes that you want to explore PostgreSQL's native implementation of replication, which is probably one of the easier to understand and implement. I belive it was a project named "DBMirror" that was moved into the official PostgreSQL Contrib beginning with 7.3.1. There are all sorts of more sophisticated replication packages out there, if you need something more feature rich:

1. Slony-1
2. PGreplication
3. PGCluster
4. PGPool
5. DBBalancer
6. Open Minds (Commercial)

user# mkdir cluster # Make this wherever you want it
user# cd cluster
    	

root# emerge --sync
root# emerge -a postgresql  #Answer "N" if you already have the latest
root# emerge -a pgperl
    	

user# equery files postgresql | grep \
	-e DBMirror \
	-e MirrorSetup \
	-e AddTrigger \
	-e slaveDatabase \
	-e pending.so
	
# Should return a list like this:
/usr/bin/DBMirror.pl
/usr/lib/postgresql/pending.so  # This is what provides the love
/usr/share/postgresql/contrib/AddTrigger.sql
/usr/share/postgresql/contrib/MirrorSetup.sql
/usr/share/postgresql/contrib/slaveDatabase.conf

    	

Tip

If you don't get this list of files Stop! You don't have everything you need. If you happen to be running another distribution you might need to install the contrib package. For example on Debian you should: apt-get install postgresql-contrib-8.x

1. 
   user# createdb -U postgres -E UTF-8 -O postgres cluster-master
   user# createdb -U postgres -E UTF-8 -O postgres cluster-slave01
   user# createdb -U postgres -E UTF-8 -O postgres cluster-slave02
       			

2. Create a simple schema file for use with populating all three:

   
   user# nano -w cluster.sql
   
   # Make the file have this content:
   CREATE TABLE car(
   	uid SERIAL PRIMARY KEY,
   	name VARCHAR,
   	make VARCHAR,
   	model VARCHAR
   );
       			

3. Populate the schema for all of the clustered db(s):

   
   user# psql -U postgres cluster-master < cluster.sql
   user# psql -U postgres cluster-slave01 < cluster.sql
   user# psql -U postgres cluster-slave02 < cluster.sql
       			

user# psql -U postgres cluster-master < /usr/share/postgresql/contrib/MirrorSetup.sql
    	

user# cp /usr/share/postgresql/contrib/AddTrigger.sql AddTrigger.sql
user# nano -w AddTrigger.sql

# Make it look like this
-- Adjust this setting to control where the objects get created.
SET search_path = public;

CREATE TRIGGER "mirror_car"
AFTER INSERT OR DELETE OR UPDATE ON "car"
FOR EACH ROW EXECUTE PROCEDURE "recordchange" ();

--Uncomment this if you also want to replicate the "boat" table:
--CREATE TRIGGER "mirror_boat"
--AFTER INSERT OR DELETE OR UPDATE ON "boat"
--FOR EACH ROW EXECUTE PROCEDURE "recordchange" ();
    	

user# psql -U postgres cluster-master < AddTrigger.sql
    	

user# cp /usr/share/postgresql/contrib/slaveDatabase.conf slave01.conf
user# cp /usr/share/postgresql/contrib/slaveDatabase.conf slave02.conf    	
    	

#########################################################################
# Config file for DBMirror.pl
# This file contains a sample configuration file for DBMirror.pl
# It contains configuration information to mirror data from 
# the master database to a single slave system.
#
# $PostgreSQL: pgsql/contrib/dbmirror/slaveDatabase.conf,v 1.3 2004/09/10 04:31:06 neilc Exp $
#######################################################################

$masterHost = "localhost";
$masterDb = "cluster-master";
$masterUser = "postgres";
$masterPassword = "postgrespassword";

# Where to email Error messages to
# $errorEmailAddr = "me@mydomain.com";

$slaveInfo->{"slaveName"} = "db-slave01";
$slaveInfo->{"slaveHost"} = "localhost";
$slaveInfo->{"slaveDb"} = "cluster-slave01";
$slaveInfo->{"slavePort"} = 5432;
$slaveInfo->{"slaveUser"} = "postgres";
$slaveInfo->{"slavePassword"} = "postgrespassword";
# If uncommented then text files with SQL statements are generated instead
# of connecting to the slave database directly. 
# slaveDb should then be commented out.
# $slaveInfo{"TransactionFileDirectory"} = '/tmp';

#
# The number of seconds dbmirror should sleep for between checking to see
# if more data is ready to be mirrored.
$sleepInterval = 60;
    	

user# psql -U postgres cluster-master

-- We're now in the psql subsystem
cluster-master=# INSERT INTO dbmirror_mirrorhost VALUES (1, 'db-slave01');
INSERT 0 1
cluster-master=# INSERT INTO dbmirror_mirrorhost VALUES (2, 'db-slave02');
INSERT 0 1

# Exit the psql subsystem:
cluster-master=# \q
    	

user# /usr/bin/DBMirror.pl slave01.conf >/dev/null 2>/dev/null &
user# /usr/bin/DBMirror.pl slave02.conf >/dev/null 2>/dev/null &
    	

# Validate all three db's are in fact empty:
user# psql -U postgres cluster-master -c 'SELECT * FROM car;'

 uid | name | make | model 
-----+------+------+-------
(0 rows)

user# psql -U postgres cluster-slave01 -c 'SELECT * FROM car;'

 uid | name | make | model 
-----+------+------+-------
(0 rows)

user# psql -U postgres cluster-slave02 -c 'SELECT * FROM car;'

 uid | name | make | model 
-----+------+------+-------
(0 rows)


# Insert a new car into our master database:
user# psql -U postgres cluster-master -c "INSERT INTO car(name, make, model) VALUES('foo', 'Honda', 'Accord');"
INSERT 0 1

# See what's in the first slave:
user# psql -U postgres cluster-slave01 -c 'SELECT * FROM car;'

 uid | name | make  | model  
-----+------+-------+--------
   1 | foo  | Honda | Accord
(1 row)


# Sweet mother!, is it in slave02 too?:
user# psql -U postgres cluster-slave02 -c 'SELECT * FROM car;'

 uid | name | make  | model  
-----+------+-------+--------
   1 | foo  | Honda | Accord
(1 row)

    	

When I get time I'll flesh this out in more detail