3.2 Clustering - Add nodes
Your initial installation is a cluster of one, a single node. Now, repeat the installation process to create three or more nodes, always ending with an odd number of nodes, as described in the Standard Deployment for guidance.
Once you have a set of nodes, the next step is to cluster them together.
CAUTION:Before proceeding with clustering, be aware that:
-
Before adding a node to a cluster, all nodes must be in a health state.
-
The node that joins a cluster loses its own application data, such as configured sessions. The data present on the node that you are joining is inherited.
-
Before joining a node to a cluster, the node must have the same products installed as those nodes that already participate in the cluster.
-
Removing a node from a cluster results in its data being lost.
-
Removing one healthy node from the cluster results in data loss and the need for a reset, but the remaining node will remain functional.
-
When later replacing a node in a cluster, always remove the existing node before adding a new node.
Follow the clustering steps for your method of deployment: Appliance or Linux installer.