Virtual Machine and Workload Management
Workload management covers AHV day-two operations: creating and cloning VMs, organizing them, keeping them available through maintenance, and moving workloads between platforms.
VM lifecycle and organization
- Images are the source for disks; clones can be full or fast (redirect-on-write). Templates capture VM configuration for repeatable builds.
- Categories drive everything downstream: policies, protection, microsegmentation and projects. Projects add quotas and scope resources to a team.
- Nutanix Guest Tools installs VirtIO drivers and enables VSS snapshots, file-level restore and self-service restore.
Availability
- AHV HA restarts VMs on surviving hosts; HA reservation reserves capacity so those restarts can actually be placed.
- Affinity policies include VM-VM affinity/anti-affinity and VM-host affinity (for licensing or performance). They constrain live migration.
- Maintenance mode evacuates a host using live migration; if anti-affinity rules or reservations block placement, the host will not enter maintenance mode.
Worked example
A two-node database pair must never share a host. Create a VM-VM anti-affinity policy containing both VMs. Later, patching host A: evacuation tries to move DB-1 to host B, where DB-2 already runs, violating the rule. The correct fix is to add a third host (or apply a documented temporary exception), not to let the rule be silently broken.
Traps
- Enabling HA without reservation and assuming VMs restart after a host loss.
- Applying anti-affinity in a two-node cluster and then expecting maintenance mode to succeed.
- Skipping NGT and then wondering why snapshots are not application-consistent.
- Overcommitting memory so heavily that HA admission or performance suffers.