PhenixID server allows for multi master cluster scenarios. General idea is that nodes in a cluster is to be considered a single entity. Each node may ha different features installed depending on the requirements. There are basically two types of set up possible when configuring clustering. Chose one or mix the two depending on load and requirements.
Regardless of cluster configuration it is key for functionality that connectivity between nodes are consistent.
Please contact support before setting up a cluster.
High availability cluster
Every node is identical. Data is shared among all nodes. All nodes can perform the same tasks.
This option is common but in some cases not ideal. Think about what is needed in the cluster. Does the configuration application really need to be present everywhere? Perhaps some functionality can be removed from one or more nodes. Deployed but unused functionality consumes resources. It is always a good practice to turn off unused features.
Scale out cluster
For deployment scenarios with heavy load scale out cluster may be a good idea. This means installing features on dedicated nodes.
For instance moving user validation to dedicated nodes may improve performance.