What crucial step must be taken before successfully stopping a Nutanix cluster?

All guest virtual machines must be shut down before stopping a Nutanix cluster. A cluster shutdown terminates the AOS services provided by the Controller VMs and ultimately shuts down the hypervisor hosts. Leaving user workloads running would expose those VMs to abrupt power loss, interrupted write operations and possible application or filesystem corruption.

The supported sequence begins with a controlled shutdown of the guest operating systems. Administrators should verify that no user VMs remain powered on, check cluster health and then execute the supported cluster-stop procedure. The cluster process coordinates the shutdown of AOS services and Controller VMs; therefore, administrators must not manually power off every CVM as the first action.

Disabling the Remote Support Tunnel has no bearing on the safe shutdown of storage and compute services. Protection domains generally do not need to be disabled for an ordinary cluster shutdown, although synchronous-replication configurations can require additional site-specific preparation.

NHCF/NCA reference: Cluster maintenance, controlled shutdown and service-management procedures; NCA Section 3, “Identify hardware maintenance tasks.” Official references: AHV—Shutting Down a Cluster , Shutting Down DR Clusters .

===============