12.2.1. Fail-Over clusters

The aim of failover clustering is to ensure that the service is accessible even if one of the servers breaks down (for example, because of a hardware error). In failover clusters only one of the nodes is functioning and carries the whole traffic, the other(s) only monitor the active node. In case of system failure resulting in a loss of service, the service is started on the other node in the system. In other words, if the active component dies the other node takes over all the services.

Note

The service fails over to the other component only in case of hardware failure, if you stop PNS no backup mechanism is initiated automatically.

Monitoring between the cluster nodes is realized with the help of Keepalived's VRRP health checks. For more information, see Section 12.5, Keepalived for High Availability.

The transfer of services can be realized using one of the following methods:

  • Transferring the Service IP address

  • Transferring IP and MAC address

  • Sending RIP messages