The master host of a VMware vSphere® High Availability cluster is responsible for detecting the failure of subordinate hosts. Depending on the type of failure detected, the virtual machines running on the hosts might need to be failed over.
In a vSphere HA cluster, three types of host failure are detected:
Failure. A host stops functioning.
Isolation. A host becomes network isolated.
Partition. A host loses network connectivity with the master host.
The master host monitors the liveness of the subordinate hosts in the cluster. This communication happens through the exchange of network heartbeats every second. When the master host stops receiving these heartbeats from a subordinate host, it checks for host liveness before declaring the host failed. The liveness check that the master host performs is to determine whether the subordinate host is exchanging heartbeats with one of the datastores. See Datastore Heartbeating. Also, the master host checks whether the host responds to ICMP pings sent to its management IP addresses.
If a master host cannot communicate directly with the agent on a subordinate host, the subordinate host does not respond to ICMP pings. If the agent is not issuing heartbeats, it is viewed as failed. The host's virtual machines are restarted on alternate hosts. If such a subordinate host is exchanging heartbeats with a datastore, the master host assumes that the subordinate host is in a network partition or is network isolated. So, the master host continues to monitor the host and its virtual machines. See Network Partitions.
Host network isolation occurs when a host is still running, but it can no longer observe traffic from vSphere HA agents on the management network. If a host stops observing this traffic, it attempts to ping the cluster isolation addresses. If this pinging also fails, the host declares that it is isolated from the network.
The master host monitors the virtual machines that are running on an isolated host. If the master host observes that the VMs power off, and the master host is responsible for the VMs, it restarts them.
If you ensure that the network infrastructure is sufficiently redundant and that at least one network path is always available, host network isolation is less likely to occur.
Proactive HA Failures
A Proactive HA failure occurs when a host component fails, which results in a loss of redundancy or a noncatastrophic failure. However, the functional behavior of the VMs residing on the host is not yet affected. For example, if a power supply on the host fails, but other power supplies are available, that is a Proactive HA failure.
If a Proactive HA failure occurs, you can automate the remediation action taken in the vSphere Availability section of the vSphere Client. The VMs on the affected host can be evacuated to other hosts and the host is either placed in Quarantine mode or Maintenance mode.
Your cluster must use vSphere DRS for the Proactive HA failure monitoring to work.