Which event triggers a failover in a Cluster Vault?

A failover in a Cluster Vault is primarily triggered by the loss of quorum ownership, which is essential for maintaining the cluster's operational integrity. Quorum is the minimum number of nodes that must be operational to ensure that the cluster can continue to function normally and make decisions.

When socket connectivity or communication among the nodes in the cluster fails, and the designated quorum is lost, the cluster cannot reach a consensus on which node should remain active. To prevent potential data corruption or inconsistency, the system initiates a failover. This ensures that operations can continue without disruption by designating another node as the primary point of access.

Other events like database updates, user authentication failures, and software updates do not specifically trigger a failover mechanism. Database updates can occur seamlessly within the active nodes, and user authentication failures typically relate to application-level issues rather than triggering a cluster-level response. Software updates are managed in a controlled manner to avoid a failure scenario and would generally be handled with maintenance procedures that do not result in a failover. Thus, the level of criticality and direct impact on the cluster's operational state makes loss of quorum ownership the primary event causing a failover.