As repmgr is a client/server PostgreSQL management suite, we need at least two nodes involved before we're really using it. We can perform the tasks outlined in this recipe as many times as we wish, creating several clones and registering them with repmgr. Of course, this book is for demonstration purposes, so we'll leave the larger clusters to you. With multiple nodes involved, the chances of data loss or system outages decline, which is excellent for our goal of high availability.

This recipe will focus on the process necessary to add a node to an existing repmgr cluster. The existing cluster in our case is the one that we established on pg-primary in the previous recipe.