What action is performed in Azure when an instance is report as being in a terminated state, but NS2 expects it to be in a running state ?

Nutanix Clusters (NC2) on Azure includes an automated host remediation feature. The NC2 portal continuously monitors the health of the Azure bare-metal instances that serve as cluster nodes. If an instance is unexpectedly terminated, the portal detects this as a host failure. It then automatically condemns the failed node, removes it from the cluster, and triggers the provisioning of a new bare-metal instance from Azure. This new instance is then automatically added to the cluster to restore its capacity and resilience, a process that requires no manual administrator intervention.

A. NS2 (NC2) operates on top of the Azure infrastructure and lacks the privileges to restart a terminated bare-metal instance directly from the Azure control plane.

B. While an alert is generated, the system is designed for automated remediation. Requiring manual replacement would defeat a key high-availability feature of the platform.

C. A terminated instance is powered off and deallocated within Azure, making it impossible for the cluster to reconnect with it over the network.

1. Nutanix. (2023). Nutanix Clusters on Microsoft Azure Deployment and User Guide v6.7. Nutanix Portal. Section: "Host Management in Nutanix Clusters on Azure", Subsection: "Host Failure and Replacement".

This section explicitly states, "If a node fails, Nutanix Clusters automatically detects the failure, removes the failed node from the cluster, provisions a new node from the Azure bare-metal inventory, and adds the new node to the cluster." This process is initiated when the portal service condemns the unhealthy node.

2. Nutanix. (2023). Nutanix Clusters on Azure Tech Note. Nutanix Portal. Section: "High Availability and Resiliency".

The document details that the NC2 control plane integrates with Azure APIs to manage the lifecycle of the bare-metal nodes. A terminated node is treated as a permanent failure, triggering the automated replacement workflow to maintain the cluster's desired node count and health status.