A three-tier application runs on AHV VMs replicated to a second Nutanix cluster. During a planned failover, the database VM must power on before the application VMs, and a runbook should execute after boot. Which Nutanix DR object should an admin use to orchestrate this multi-VM failover?

A recovery plan is the Nutanix Disaster Recovery object in Prism Central that turns replicated VMs into an ordered failover sequence. It works with VMs protected by a protection policy, maps networks between sites, sets the power-on order through stages with optional delays, and can run in-guest scripts so post-boot tasks execute when the plan runs. That makes it the right choice when several VMs must come up as a service, such as the database before the application tier, rather than failing over in an arbitrary order. A protection domain is the legacy Prism Element construct that groups VMs for snapshots and replication on a schedule; it decides what is protected together but doesn't define startup dependencies or post-boot automation. An async replication schedule controls how often recovery points are sent to the remote site and helps meet an RPO, yet it doesn't manage failover sequencing. A NearSync schedule tightens the RPO to between 1 and 15 minutes, but it is still a replication timing setting, not an orchestration layer for multi-VM recovery. Exam caveat: separate the object that protects data (protection policy or protection domain) from the object that orchestrates failover and failback (recovery plan). Operational check: review the recovery plan's stages, network mappings and scripts, then run a test failover before the planned event.