As a prerequisite to running Slony-I on PostgreSQL, the slony modules xxid and slony1_funcs must be present on all servers that have to run a Slony-I replication node. This is usually done by the Slony-I installation routine.
Setting up Slony-I for the first time can be a daunting task. The following sections of the pgAdmin helpfile will help guide you through creation of your first cluster.
Note At this point, the slon daemons (service engines on Windows) should be started.
To install a Slony-I cluster on the first database, the “New Slony-I Cluster” dialog is used. It executes the official Slony-I cluster creation SQL scripts, which are located in the directory configured in the Options dialog.
pgAdmin III needs to store information how to contact each individual node in the cluster. To achieve this, pgAdmin III uses the concept of “Administrative nodes”.
After the first node in the Slony-I replication cluster has been successfully created, all subsequent nodes take their configuration and procedures from the first nodes. This process is called “Joining a cluster” in pgAdmin III. Usually, you should also select an existing node as admin node, to insure later accessibility from pgAdmin III.
After you added a new node to the Slony-I cluster, you need to set up replication paths between the nodes, to enable communication between the nodes.
When a cluster is to be upgraded to a new version of the Slony-I clustering software, the upgrade process has to be run on all nodes of the cluster. For each node, the slon daemon needs to be stopped, then the upgrade dialog is started and a node with the new software is selected (pgAdmin III will extract all software from that node), and finally the slon daemon is started again.
Currently, pgAdmin III does not support upgrading from slony creation scripts. Instead, create an intermediate cluster from the creation scripts, use it as a source for the upgrade dialog, and drop the cluster after usage. You may also use the slonik tool to upgrade the first node, and then use it as source for subsequent node upgrades.