VMware HA is an automated restart feature for VMs that are running in a cluster of ESX hosts. HA monitors hosts and VMs and uses a heartbeat to minimize downtime. As with DRS, shared storage is required for the HA feature.

When HA detects a failure, it switches the unresponsive VM to a new host and restarts it. Because the running state of the VM is not preserved, an HA-switched IP Availability Manager VM would be restarted from the last saved repository (RPS) file. That file might be several hours old, and restoring that file might take a significant length of time, especially in the case of a large managed network topology.