Postgres-XC allows you to add new servers to the setup at any point in the process. All you have to do is to set up a node as we have shown before and call CREATE NODE on the controller. The system will then be able to use this node.

However, there is one important thing about this. If you have partitioned a table before adding a new node, this partitioned table will stay at its place. Some people would expect that Postgres-XC magically rebalances this data to new nodes. This is not going to happen. It is your task to move new data there and make good use of the server.

It is necessary for Postgres-XC to behave that way because otherwise adding a new node would need locking up existing infrastructure to rebalance data. Doing that is clearly not acceptable.

If you have to extend your cluster by many machines, it can be beneficial to recreate a table using new rules. Doing this without downtime is not trivial, and you have to come up with a strategy serving your needs. At the time this book was written, there was no out-of-the-box solution.