Virtualization and Cloud Infrastructure
What this domain tests
Virtualization and Cloud Infrastructure tests AHV VM lifecycle, cluster placement, images, networks, affinity, HA, live migration, Prism management, and hybrid-cloud extension. You need to reason about capacity and availability, not just create VMs.
Core rules
- A cluster must retain enough CPU and memory for the failure it is designed to absorb. HA reservation, N+1, and overcommit interact; high consolidation can make HA ineffective.
- Affinity and anti-affinity rules control placement. Anti-affinity protects paired workloads from the same host failure; affinity can reduce cross-host traffic but adds risk.
- Use categories and projects for scoping, policy, and automation. Avoid one-off VM names as the only inventory.
- Images, networks, and storage containers should be standardized before scaling. Live migration depends on compatible CPU features and network reachability.
Worked example
Three hosts each have 256 GB RAM. Total VM memory is 600 GB. With N+1, two hosts must carry all VMs after one failure: 512 GB available, so the design fails. Reduce VM memory to 480 GB, or add a host, or accept reduced HA. Document the chosen trade-off. Then check CPU: if total vCPU demand exceeds two hosts' physical cores after overcommit limits, memory alone is not enough.
Traps
Counting total cluster capacity instead of post-failure capacity; ignoring CPU ready and overcommit; applying anti-affinity that makes HA impossible; mismatched CPU features blocking live migration; and treating cloud extension as unlimited capacity.