AHV Virtualization and VM Lifecycle Management
The AHV model
AHV is Nutanix's native KVM-based hypervisor, managed by the Acropolis service inside the CVM. There is no external vCenter and no datastore to format: a VM's disks are vDisks in a storage container, thinly provisioned and reachable from every host. That is why live migration works without shared external storage or a dedicated migration network - the data is already cluster-wide.
Lifecycle
Typical flow: upload or import an image, create a VM (vCPU, memory, disks, NIC on a VLAN), power it on, then clone, snapshot, or migrate as needed. VM HA restarts a VM on a surviving host when its host fails; it is a restart, not a live migration, and it depends on HA capacity being reserved in the cluster. Affinity and anti-affinity policies keep chosen VMs together or apart.
Guest integration
Nutanix Guest Tools (NGT) is installed inside the guest and enables VSS-consistent snapshots, self-service restore, and file-level restore. It is per-VM and does nothing until installed.
Traps
Expect distractors that describe vSphere concepts - resource pools, datastores, vMotion networks - as if they applied to AHV. Also remember that HA cannot protect VMs if the cluster lacks reserved capacity, and that deleting a VM does not necessarily delete its snapshots unless you choose to.
Rule of thumb
Lifecycle questions reduce to: image, then VM, then power on, then protect (snapshot, HA, NGT). An answer that skips NGT for a consistency question is usually wrong.