Data Protection, Disaster Recovery, and Migration
Protection primitives
Nutanix data protection uses protection domains, snapshots, and replication. A protection domain groups VMs and schedules local snapshots and remote replication. Replication can be asynchronous or near-sync, with RPO set by schedule and bandwidth. Replica clusters need enough capacity and network path. RF2 tolerates one node failure in a three-node cluster; RF1 on a single node does not.
DR and migration
- Nutanix Leap orchestrates DR runbooks: planned failover, unplanned failover, and failback. It depends on working replication and DNS/network changes.
- Nutanix Move migrates VMs from supported sources, including AWS EC2 and VMware, to AHV. Install the Move appliance, create a plan, test, then cut over.
- For NC2, replicate from on-premises to the AWS cluster over VPN or Direct Connect; validate MTU and throughput before relying on the RPO.
Worked example
An app has a 1-hour RPO. Create a protection domain containing the app VMs, schedule snapshots every hour, and replicate to an NC2 cluster. Leap maps the VMs to the recovery cluster and a runbook powers them on in the right order. During a test, the replica is two hours behind because the VPN is saturated; the RPO is not met, so the fix is bandwidth or a change to near-sync, not a shorter snapshot schedule alone.
Traps
Treating snapshots as offsite backups; forgetting that replication targets need capacity; assuming Leap fixes broken replication; testing failover without testing failback; using RF1 for production.