NCP-BC-7.5 Exam Dumps - Nutanix Certified Professional - Business Continuity (NCP-BC) 7.5

Nutanix manages data recovery operations differently depending on whether a virtual machine is in an asynchronous or synchronous replication state. For a standard Asynchronous VM like " REPORTING-DEV, " the local cluster takes periodic snapshots. If data corruption occurs, the administrator can perform an " In-place Restore " or " Revert " using a local recovery point. This is a standard procedure where the current vDisks are overwritten by the point-in-time data from the snapshot.

However, " ERP-PROD " is part of a Synchronous replication policy (0 RPO). This means the storage container it resides on is effectively " stretched " and locked in an active-active relationship with a remote cluster. Every write must be identical at both sites. If an administrator attempts to " Revert " a synchronous VM to a local snapshot, it would cause an immediate data mismatch between the primary and recovery sites, breaking the zero-RPO guarantee and potentially causing cluster-wide consistency issues. To protect data integrity, Prism Central typically grays out or blocks the " Revert " option for VMs currently in an active synchronous " Sync " state. To recover such a VM, an administrator would generally need to either " Clone " the snapshot to a new VM or temporarily move the VM into an asynchronous state by modifying the policy, thereby unlocking the ability to revert the data disks locally before re-enabling synchronization.

When a primary Nutanix site suffers a catastrophic failure, the disaster recovery plan must be executed from the recovery (secondary) site . In a legacy Protection Domain-based environment, the secondary site is in a " passive " or " standby " mode, receiving snapshots but not actively running the workloads. The traditional method to failover is using the " Activate " command (Option A), which promotes the PD and registers the VMs. However, in certain specific failure scenarios or older software versions, the secondary cluster may require a service-level reset to recognize its new role as the active master for those specific workloads.

The Cerebro service is the core component responsible for managing Protection Domains and replication metadata. If the primary site is completely inaccessible and the secondary site ' s metadata state needs refreshing to initiate recovery, restarting the Cerebro service on the secondary cluster CVMs can force a re-evaluation of all local snapshots and protection domain states. This action triggers the recovery logic, allowing the administrator to then see and activate the recovery points. While " Activate " is the goal, the first technical step in a situation where the cluster management plane is unstable or the primary is gone is ensuring the data protection services are aware of the current state. This allows the VMs to be restored from the latest snapshots, fulfilling the Recovery Time Objective (RTO) despite the complete loss of the source infrastructure.

Disaster recovery orchestration in the Nutanix environment often involves the pairing of two separate Availability Zones (AZs), each managed by its own Prism Central instance. For these two management planes to communicate, exchange metadata, and synchronize Protection Policies and Recovery Plans, specific firewall ports must be open between them.

Port 9440 is the standard port used for all Prism Central and Prism Element web interface and API traffic. During the AZ pairing process, the primary Prism Central must reach the recovery Prism Central ' s API to establish the trust relationship and pair the sites. Other ports mentioned, such as 2009 (Option A) and 2020 (Option B), are used for data replication between Controller VMs (CVMs) but are not used for the initial management-plane pairing between Prism Central instances. Port 2074 (Option C) is typically used for Nutanix Guest Tools (NGT) communication with the cluster. Therefore, ensuring port 9440 is open between the two Prism Central instances is a mandatory prerequisite for successful disaster recovery site pairing.

Nutanix Cloud Clusters (NC2) on AWS or Azure allows for flexible disaster recovery models, including " Pilot Light " and " Zero Compute " . " Zero Compute " is a cost-effective model where data is replicated to cloud storage (like S3 or Azure Blob) without needing a running cluster at the recovery site. Multi-Site Tooling (MST) or Nutanix Disaster Recovery orchestration can be used to manage this process.

In a " Zero Compute " setup, the orchestration of an automated failover is technically complex because there is no active cluster management plane at the recovery site until the failover is triggered. A key technical requirement for successful automated failover in this model is the management of networking and IP addresses. Because the VMs are being brought up in a cloud environment that likely has a different network subnet than the on-premises environment, the system must be able to dynamically assign and manage IP addresses. Therefore, configuring an external IPAM (IP Address Management) system is a strict requirement to ensure that when VMs are restored, they receive valid networking configurations that allow them to communicate with external users and other services. Without a properly configured IPAM and network mapping in the Recovery Plan, the restored workloads would be isolated and non-functional. This highlights the need for careful network planning when designing DR strategies that utilize cloud-native storage and on-demand compute resources.

Virtual GPUs (vGPUs) are hardware-dependent resources that are tied to specific physical GPU cards installed in the Nutanix nodes. While Nutanix Disaster Recovery can replicate the VM ' s virtual disks and general configuration, the specific mapping to a physical vGPU profile is often not automatically preserved across clusters due to potential differences in hardware availability or GPU generations at the recovery site.

When the VM fails over and boots up at the secondary site, the guest OS may see the NVIDIA driver but will find that the " backed " hardware resource is missing or incorrectly mapped, leading to a failure in hardware acceleration. To resolve this, the administrator must manually edit the VM ' s hardware settings at the recovery site and re-assign a compatible vGPU profile from the local cluster ' s available GPU resources. This post-failover cleanup task is essential for workloads like VDI (Virtual Desktop Infrastructure) or CAD applications that depend on GPU processing. Relying on NGT (Option A) or reinstalling the OS (Option B) will not fix the underlying missing hardware assignment in the hypervisor, highlighting the need for specialized knowledge when protecting VMs with specialized hardware pass-through or vGPU requirements.