Hello Charu,

There are two scenarios, first one occur when node failed and VMs restarted on another node but there is no available resources after another failure, then the HA status of the cluster goes into Critical state. This happens because the cluster cannot tolerate any more host failures. You have to ensure that you bring back the host so that your cluster is protected against any further host failures.

Second one on a less loaded cluster, it is possible for HA to go directly back to OK state if enough resources is reserved to protect another host failure. The start and migrate operations on the VMs are restricted in the Critical state because Acropolis continuously tries to ensure that the HA status is back to the OK state.

HA enabled setting is to ensure incase of any node failure VMs will be rebooted on available hosts.

if HA is disabled no reservations of node or memory on the node are done in the cluster, and in case of any node failures, the virtual machines are restarted on the remaining nodes based on the resources/memory available on the node.