PostgreSQL Database Replication

posted Sep 17, 2010, 2:23 PM by Sachchida Ojha
In this example, we will be replicating a brand new pgbench database. The mechanics of replicating an existing database are covered here, however we recommend that you learn how Slony-I functions by using a fresh new non-production database.

Note that pgbench is a "benchmark" tool that is in the PostgreSQL set of contrib tools. If you build PostgreSQL from source, you can readily head to contrib/pgbench and do a make install to build and install it; you may discover that included in packaged binary PostgreSQL installations.

The Slony-I replication engine is trigger-based, allowing us to replicate databases (or portions thereof) running under the same postmaster.

This example will show how to replicate the pgbench database running on localhost (master) to the pgbench slave database also running on localhost (slave). We make a couple of assumptions about your PostgreSQL configuration:

- You have tcpip_socket=true in your postgresql.conf and
- You have enabled access in your cluster(s) via pg_hba.conf

The REPLICATIONUSER needs to be a PostgreSQL superuser. This is typically postgres or pgsql, although in complex environments it is quite likely a good idea to define a slony user to distinguish between the roles.

You should also set the following shell variables:


Here are a couple of examples for setting variables in common shells:

bash, sh, ksh export CLUSTERNAME=slony_example
(t)csh: setenv CLUSTERNAME slony_example

If you're changing these variables to use different hosts for MASTERHOST and SLAVEHOST, be sure not to use localhost for either of them. This will result in an error similar to the following:

ERROR remoteListenThread_1: db_getLocalNodeId() returned 2 - wrong database?
1.1. Creating the pgbench user

createuser -A -D $PGBENCHUSER
1.2. Preparing the databases


Because Slony-I depends on the databases having the pl/pgSQL procedural language installed, we better install it now. It is possible that you have installed pl/pgSQL into the template1 database in which case you can skip this step because it's already installed into the $MASTERDBNAME.

createlang -h $MASTERHOST plpgsql $MASTERDBNAME

Slony-I does not automatically copy table definitions from a master when a slave subscribes to it, so we need to import this data. We do this with pg_dump.


To illustrate how Slony-I allows for on the fly replication subscription, let's start up pgbench. If you run the pgbench application in the foreground of a separate terminal window, you can stop and restart it with different parameters at any time. You'll need to re-export the variables again so they are available in this session as well.

The typical command to run pgbench would look like:

pgbench -s 1 -c 5 -t 1000 -U $PGBENCHUSER -h $MASTERHOST $MASTERDBNAME

This will run pgbench with 5 concurrent clients each processing 1000 transactions against the pgbench database running on localhost as the pgbench user.

1.3. Configuring the Database for Replication.

Creating the configuration tables, stored procedures, triggers and configuration is all done through the slonik(1) tool. It is a specialized scripting aid that mostly calls stored procedures in the master/slave (node) databases. The script to create the initial configuration for the simple master-slave setup of our pgbench database looks like this:

#!/bin/sh          slonik <<_EOF_         #--         # define the namespace the replication system uses in our example it is         # slony_example         #--         cluster name = $CLUSTERNAME;              #--         # admin conninfo's are used by slonik to connect to the nodes one for each         # node on each side of the cluster, the syntax is that of PQconnectdb in         # the C-API         # --         node 1 admin conninfo = 'dbname=$MASTERDBNAME host=$MASTERHOST user=$REPLICATIONUSER';         node 2 admin conninfo = 'dbname=$SLAVEDBNAME host=$SLAVEHOST user=$REPLICATIONUSER';              #--         # init the first node.  Its id MUST be 1.  This creates the schema         # _$CLUSTERNAME containing all replication system specific database         # objects.              #--         init cluster ( id=1, comment = 'Master Node');               #--         # Because the history table does not have a primary key or other unique         # constraint that could be used to identify a row, we need to add one.         # The following command adds a bigint column named         # _Slony-I_$CLUSTERNAME_rowID to the table.  It will have a default value         # of nextval('_$CLUSTERNAME.s1_rowid_seq'), and have UNIQUE and NOT NULL         # constraints applied.  All existing rows will be initialized with a         # number         #--         table add key (node id = 1, fully qualified name = 'public.history');              #--         # Slony-I organizes tables into sets.  The smallest unit a node can         # subscribe is a set.  The following commands create one set containing         # all 4 pgbench tables.  The master or origin of the set is node 1.         #--         create set (id=1, origin=1, comment='All pgbench tables');         set add table (set id=1, origin=1, id=1, fully qualified name = 'public.accounts', comment='accounts table');         set add table (set id=1, origin=1, id=2, fully qualified name = 'public.branches', comment='branches table');         set add table (set id=1, origin=1, id=3, fully qualified name = 'public.tellers', comment='tellers table');         set add table (set id=1, origin=1, id=4, fully qualified name = 'public.history', comment='history table', key = serial);              #--         # Create the second node (the slave) tell the 2 nodes how to connect to         # each other and how they should listen for events.         #--              store node (id=2, comment = 'Slave node');         store path (server = 1, client = 2, conninfo='dbname=$MASTERDBNAME host=$MASTERHOST user=$REPLICATIONUSER');         store path (server = 2, client = 1, conninfo='dbname=$SLAVEDBNAME host=$SLAVEHOST user=$REPLICATIONUSER');         store listen (origin=1, provider = 1, receiver =2);         store listen (origin=2, provider = 2, receiver =1);     _EOF_


Is the pgbench still running? If not, then start it again.

At this point we have 2 databases that are fully prepared. One is the master database in which pgbench is busy accessing and changing rows. It's now time to start the replication daemons.

On $MASTERHOST the command to start the replication engine is


Likewise we start the replication system on node 2 (the slave)


Even though we have the slon(1) running on both the master and slave, and they are both spitting out diagnostics and other messages, we aren't replicating any data yet. The notices you are seeing is the synchronization of cluster configurations between the 2 slon(1) processes.

To start replicating the 4 pgbench tables (set 1) from the master (node id 1) the the slave (node id 2), execute the following script.

    #!/bin/sh     slonik <<_EOF_          # ----          # This defines which namespace the replication system uses          # ----          cluster name = $CLUSTERNAME;               # ----          # Admin conninfo's are used by the slonik program to connect          # to the node databases.  So these are the PQconnectdb arguments          # that connect from the administrators workstation (where          # slonik is executed).          # ----          node 1 admin conninfo = 'dbname=$MASTERDBNAME host=$MASTERHOST user=$REPLICATIONUSER';          node 2 admin conninfo = 'dbname=$SLAVEDBNAME host=$SLAVEHOST user=$REPLICATIONUSER';               # ----          # Node 2 subscribes set 1          # ----          subscribe set ( id = 1, provider = 1, receiver = 2, forward = no);     _EOF_

Any second now, the replication daemon on $SLAVEHOST will start to copy the current content of all 4 replicated tables. While doing so, of course, the pgbench application will continue to modify the database. When the copy process is finished, the replication daemon on $SLAVEHOST will start to catch up by applying the accumulated replication log. It will do this in little steps, 10 seconds worth of application work at a time. Depending on the performance of the two systems involved, the sizing of the two databases, the actual transaction load and how well the two databases are tuned and maintained, this catchup process can be a matter of minutes, hours, or eons.

You have now successfully set up your first basic master/slave replication system, and the 2 databases should, once the slave has caught up, contain identical data. That's the theory, at least. In practice, it's good to build confidence by verifying that the datasets are in fact the same.

The following script will create ordered dumps of the 2 databases and compare them. Make sure that pgbench has completed, so that there are no new updates hitting the origin node, and that your slon sessions have caught up.

    #!/bin/sh     echo -n "**** comparing sample1 ... "     psql -U $REPLICATIONUSER -h $MASTERHOST $MASTERDBNAME >dump.tmp.1.$$ <<_EOF_          select 'accounts:'::text, aid, bid, abalance, filler               from accounts order by aid;          select 'branches:'::text, bid, bbalance, filler               from branches order by bid;          select 'tellers:'::text, tid, bid, tbalance, filler               from tellers order by tid;          select 'history:'::text, tid, bid, aid, delta, mtime, filler,               "_Slony-I_${CLUSTERNAME}_rowID"               from history order by "_Slony-I_${CLUSTERNAME}_rowID";     _EOF_     psql -U $REPLICATIONUSER -h $SLAVEHOST $SLAVEDBNAME >dump.tmp.2.$$ <<_EOF_          select 'accounts:'::text, aid, bid, abalance, filler               from accounts order by aid;          select 'branches:'::text, bid, bbalance, filler               from branches order by bid;          select 'tellers:'::text, tid, bid, tbalance, filler               from tellers order by tid;          select 'history:'::text, tid, bid, aid, delta, mtime, filler,               "_Slony-I_${CLUSTERNAME}_rowID"               from history order by "_Slony-I_${CLUSTERNAME}_rowID";     _EOF_          if diff dump.tmp.1.$$ dump.tmp.2.$$ >$CLUSTERNAME.diff ; then          echo "success - databases are equal."          rm dump.tmp.?.$$          rm $CLUSTERNAME.diff     else          echo "FAILED - see $CLUSTERNAME.diff for database differences"     fi

Note that there is somewhat more sophisticated documentation of the process in the Slony-I source code tree in a file called slony-I-basic-mstr-slv.txt.

If this script returns FAILED please contact the developers at