Because cluster resource services uses only IP to communicate with other cluster nodes, all cluster nodes must be IP-reachable.
This means that you must have IP interfaces configured to connect the nodes in your cluster. These IP addresses must be set up either manually by the network administrator in the TCP/IP routing tables on each cluster node or they may be generated by routing protocols running on the routers in the network. This TCP/IP routing table is the map that clustering uses to find each node; therefore, each node must have its own unique IP address. Each node may have up to two IP addresses assigned to it. These addresses must not be changed in any way by other network communications applications. Be sure when you assign each address that you take into account which address uses which kind of communication line. If you have a preference for using a specific type of communication media, make sure to configure the first IP address using your preferred media. The first IP address is what is treated preferentially by the reliable message function and heartbeat monitoring. All IP addresses on a node must be able to reach every other IP address in the cluster. One address can reach another address if you can ping and use a UDP message traceroute in both directions.