Skip to content

Nutanix Data Protection and Disaster Recovery

Scope

Data protection, replication, backup, and disaster recovery for Nutanix environments: protection domains, Leap orchestration, NearSync and metro availability replication tiers, third-party backup integration, immutable backup targets (Objects with WORM), retention policies, and DR testing.

Checklist

  • [Critical] Are protection domains (PDs) configured for legacy VM-centric snapshot and replication, or has the environment migrated to Prism Central-based protection policies using Leap for orchestrated DR?
  • [Critical] Are consistency groups used within protection domains to snapshot multiple related VMs simultaneously (e.g., app server + database), ensuring crash-consistent recovery points across dependent workloads?
  • [Critical] Are snapshot schedules configured with appropriate RPO -- hourly for general production, NearSync (1-minute RPO) for critical databases, synchronous replication (metro availability) for zero-RPO applications?
  • [Critical] Is Nutanix Leap configured with recovery plans that define VM boot order, network mappings (source-to-target VLAN mapping), IP re-addressing rules, and post-recovery scripts for automated failover?
  • [Critical] Is metro availability (synchronous replication) deployed only where network latency between sites is under 5ms RTT, with a witness VM at a third site to handle split-brain scenarios?
  • [Recommended] Is NearSync replication configured for Tier-1 applications requiring 1-minute RPO, understanding that it requires AOS 5.17+ and consumes more network bandwidth and CVM resources than async replication?
  • [Recommended] Are async replication schedules configured with bandwidth throttling to prevent DR replication from saturating the WAN link and impacting production traffic?
  • [Recommended] Is third-party backup software (Veeam, HYCU, Commvault) integrated using Nutanix API v3 or Changed Block Tracking (CBT) for efficient incremental backups rather than full-image copies?
  • [Recommended] Are backup targets configured on Nutanix Objects (S3-compatible, on-cluster or separate cluster) for cost-effective, air-gapped backup storage with WORM (Write Once Read Many) for ransomware protection?
  • [Recommended] Are retention policies defined per workload tier -- 7-day local snapshots for fast restore, 30-day remote snapshots for DR, 90-day+ backup copies on Objects or external NAS for compliance?
  • [Recommended] Is DR testing performed quarterly using Leap's test failover capability, which creates an isolated network bubble to validate recovery plans without impacting production?
  • [Recommended] Are DR runbooks documented with clear RTO/RPO targets per application, failover procedures, DNS/IP switchover steps, and validated rollback procedures?
  • [Optional] Is the replication bandwidth between primary and DR sites sized for both steady-state change rate and initial seeding, with compression enabled on replication streams to reduce WAN consumption?
  • [Optional] Are Nutanix Guest Tools (NGT) installed on Windows VMs to enable application-consistent (VSS-quiesced) snapshots rather than crash-consistent snapshots?

Why This Matters

Nutanix provides multiple data protection tiers, each with different RPO, complexity, and resource costs. Protection domains with async replication are the simplest approach but provide only crash-consistent snapshots at hourly or longer intervals. NearSync reduces RPO to 1 minute but requires lightweight snapshots and additional CVM resources. Metro availability provides synchronous replication with zero RPO but requires low-latency (<5ms) links and a witness VM -- without the witness, a network partition forces manual intervention to avoid split-brain. Leap adds orchestration on top of any replication method, automating the complex sequence of powering on VMs in order, remapping networks, and re-addressing IPs -- without Leap, DR failover is a manual, error-prone process under pressure. Backup is distinct from replication: replication protects against site failure, while backup with immutable targets (Objects with WORM) protects against ransomware and accidental deletion. Most organizations need both.

Common Decisions (ADR Triggers)

  • Replication tier -- Async (hourly+ RPO, lowest resource cost) vs NearSync (1-minute RPO, moderate overhead) vs metro availability (zero RPO, requires low-latency link and witness)
  • DR orchestration -- Nutanix Leap (built-in, recovery plans, test failover) vs third-party DR orchestration (Zerto, VMware SRM with ESXi) vs manual runbooks
  • Backup solution -- HYCU (Nutanix-native, API-integrated) vs Veeam (broad ecosystem, mature) vs Commvault (enterprise scale, complex) vs Rubrik (SaaS management)
  • Backup target -- Nutanix Objects with WORM (on-cluster, immutable) vs external NAS (simple, familiar) vs cloud tier to AWS S3/Azure Blob (offsite, pay-per-use)
  • Snapshot consistency -- Crash-consistent (default, no guest dependency) vs application-consistent (requires NGT/VSS on Windows, pre/post scripts on Linux)
  • Retention strategy -- Short local + long remote (optimize local storage) vs long local + archive offsite (fastest restore) vs tiered with lifecycle policies
  • DR site architecture -- Symmetric (identical cluster at DR site, bidirectional replication) vs asymmetric (smaller DR cluster, one-way replication, accept degraded performance during failover)

See Also

  • general/disaster-recovery.md -- general DR planning (RPO/RTO, tiering)
  • general/enterprise-backup.md -- enterprise backup architecture patterns
  • providers/nutanix/infrastructure.md -- Nutanix cluster design for DR
  • providers/nutanix/storage.md -- storage replication and snapshot management