NFS server virtual IP address

A virtual IP address is required to connect NFS clients to an NFS server by mounting an NFS file system from the NFS server using the virtual IP address.

NFS clients connect to an NFS server by mounting an NFS file system from the NFS server, which is determined by the host name or the IP address of the host. When the NFS server resource is restarted on a different node, then the virtual IP address (represented by class IBM.ServiceIP) chooses a network adapter on the other node based on its attributes IPAddress and NetMask, so that the client always connects to the same IP address.

Example

node-1 
      network adapter eth0 192.168.2.2 netmask 255.255.255.0
      network adapter eth1 10.0.10.2 netmask 255.255.255.0
node-2 
      network adapter eth0 192.168.2.3 netmask 255.255.255.0
      network adapter eth1 10.0.10.3 netmask 255.255.255.0
In the above configuration, suppose that you only want NFS clients to access the nodes through IP addresses of 192.168.2.2 and 192.168.2.3. Then you must use eth0 of node-1 and node-2 in the IBM.Equivalency of IBM.NetworkInterface. The IBM.ServiceIP that depends on this equivalency can have any free IP address in the subnet. Check with your network administrator for an IP address that is available, for example 192.168.2.4: 
nfsserver-ip attributes:
	 IPAddress 192.168.2.4
	 NetMask 255.255.255.0
	 NodeNameList {"node-1","node-2"}
System Automation for Multiplatforms knows that IBM.ServiceIP depends on network interfaces eth0 of node-1 and node-2, chooses a network adapter on the node on which the NFS server runs, and passes it to IBM.ServiceIP, which creates an IP alias on that network adapter.
Note: For the setup of the high availability policy, you should:
  • Gather the network adapters of the nodes, for example using ifconfig and select the ones which should be used by NFS clients to access data through the NFS server.
  • Get a free IP address and the corresponding netmask from your network administrator.