NCP-BC-7.5 Questions and Answers

Nutanix NearSync replication provides a middle ground between traditional asynchronous replication and synchronous mirroring, offering RPOs as low as 1 minute using Lightweight Snapshots (LWS). However, the LWS mechanism and the high-frequency metadata operations required to maintain a 1-to-15 minute RPO have specific hardware and cluster-level requirements.

One of the strict prerequisites for enabling NearSync is the cluster size. Nutanix requires a minimum of a three-node cluster for NearSync replication. This is because the system must have enough resources to distribute the LWS metadata and handle the increased I/O overhead without compromising the cluster ' s availability or performance. In a two-node cluster (which is common for small ROBO deployments), the system does not meet the minimum redundancy and resource thresholds required for the NearSync engine to operate reliably. While two-node clusters support standard Asynchronous replication (with a 60-minute RPO or higher), they are restricted from using the NearSync lightweight snapshot engine. An administrator attempting to set a 5-minute RPO on a two-node cluster will find the option either grayed out or the task will fail validation. To achieve a 5-minute RPO, the organization would need to expand the remote office cluster to at least three nodes or settle for a higher RPO supported by standard asynchronous replication on the existing hardware.

Testing is the most vital component of a resilient BCDR strategy, as it proves that recovery objectives (RTO/RPO) are achievable before a real emergency occurs. The Nutanix " Test Failover " feature is purpose-built for this requirement. Unlike a Planned Failover (which shuts down production) or an Unplanned Failover (which assumes production is dead), a Test Failover performs a non-disruptive validation.

During a Test Failover, the orchestration engine uses the most recent snapshots to power on VMs at the destination cluster. To satisfy the " without impacting production " requirement, the test VMs are typically attached to an " Isolated " or " Test " network segment that has no route to the live production network. This prevents IP address conflicts and service interference while allowing the administrator to log into the recovered VMs and verify that applications start correctly. Simple " Validation " (Option A) only checks the configuration settings and connectivity but does not actually test the power-on sequence or application logic. By executing a full Test Failover, the organization gains documented proof of its DR readiness, identifying potential issues like missing dependencies or incorrect boot orders without causing any disruption to the users currently accessing the primary site.

Single Root I/O Virtualization (SR-IOV) allows a virtual machine to bypass the hypervisor ' s virtual switch and communicate directly with a physical network adapter ' s virtual functions. This is commonly used for high-performance applications that require extremely low latency and high throughput. However, SR-IOV creates a " hard " dependency between the virtual machine and the specific physical hardware (NIC) of the host it is running on.

Disaster Recovery involves moving a VM to an entirely different cluster, which likely has different physical network adapters or a different SR-IOV configuration. Nutanix Disaster Recovery orchestration is designed to ensure that VMs can always boot at the recovery site, even if specialized hardware resources are unavailable. For VMs configured with SR-IOV, the recovery process includes a built-in fallback mechanism. During failover, the system will recover the VM and power it on, but it will disable the SR-IOV attachment. To ensure connectivity, it automatically adds or falls back to a standard virtio-net virtual adapter. This allows the VM to have network access immediately, though at a potentially lower performance level. The administrator must then manually re-configure SR-IOV at the recovery site if the local hardware supports it. This design prioritizes service availability (powering on and communicating) over specialized hardware performance during a disaster recovery event.

Migrating from legacy Protection Domains to modern Prism Central Protection Policies involves transitioning the management of virtual machine snapshots. During this process, the historical snapshots created by the Protection Domain are often intended to be " claimed " or utilized by the new PC-based system as valid recovery points.

Snapshots are point-in-time references to data on the storage tier. If a snapshot is deleted from the cluster before the migration to Protection Policies is finalized, the underlying data blocks associated with that specific point in time may be reclaimed by the system ' s garbage collection process. Consequently, the recovery points that were associated with that deleted snapshot will no longer be available for use within the new deployment. The system cannot " automatically recreate " (Option A) a deleted snapshot because the exact historical state of the data is lost once the snapshot is purged. The migration itself typically continues for the remaining valid data, but the loss of historical recovery points could impact the business ' s ability to restore to specific past dates. This highlights the importance of data retention awareness during management-plane migrations.

Modern Nutanix Disaster Recovery (Leap) relies on " Categories " to automate the inclusion of virtual machines into Protection Policies and Recovery Plans. When a Protection Policy is created, it is typically applied to a category (e.g., AppType:SQL) rather than individual VM names.

In this scenario, the newly deployed VMs were missed during a test failover despite other VMs on the same cluster recovering successfully. The most likely root cause is that the application team, working independently, failed to assign the required categories to the new VMs during the provisioning process. Because the VMs lacked the correct category tags, the Nutanix Protection Policy did not " see " them, and consequently, no recovery points were ever created for them. This is a common operational failure in environments where " infrastructure as code " or strict provisioning standards are not enforced. Recovery Plans (Option A) do not need manual re-synchronization for new VMs if the categories are correctly mapped. While an initial sync is required (Option D), it would typically show a task or progress rather than just leaving the VMs " missing " from the entire protection summary. Therefore, verifying category assignment is the primary troubleshooting step.

The primary purpose of a Test Failover is to validate the recovery process in a non-disruptive manner. When a VM is recovered from a snapshot during a test, it retains its original configuration, including its IP address and hostname. If this test VM were to be powered on on the same network as the live production VM, it would immediately cause a duplicate IP address conflict and potentially disrupt production services or trigger security alerts.

To prevent this disruption, Nutanix Recovery Plans allow administrators to map the production network to a " Test Network " at the recovery site. This test network is typically an " Isolated " VLAN or overlay network that has no physical or logical route back to the production environment. By using an isolated network (Option A), the test VMs can boot up, run their internal services, and even communicate with other test VMs (if placed on the same isolated segment) without any risk of interfering with real users or databases. This allows the infrastructure team to verify that the " Recovered " state is functional while the " Primary " state continues to serve the business, fulfilling the core requirement of non-disruptive testing.

Nutanix Disaster Recovery to the Cloud is built on the principle of automation and resource optimization. When a failover to a Nutanix Cloud Availability Zone (AZ) occurs, the orchestration engine can dynamically provision virtual networks to ensure recovered VMs have immediate connectivity. These networks are temporary infrastructure components required only while the workloads are running in the cloud. Once the primary on-premises site is restored and a failback is successfully executed, the VMs are migrated back to their original home. At this point, the dynamically created virtual networks at the cloud recovery site no longer host any active entities. The Nutanix orchestration engine is designed to automatically clean up these ephemeral resources (Option B) to prevent " infrastructure sprawl " and to ensure the customer is not billed for unused networking components in the cloud. This automated cleanup is a core feature of the Nutanix Cloud AZ workflow, allowing administrators to focus on stabilizing production rather than performing manual housekeeping tasks on the recovery site. This cycle of dynamic provisioning and automated cleanup ensures that the DR environment is always " right-sized " and ready for the next event.

The Nutanix Cloud Gateway serves as a critical bridge between on-premises Nutanix clusters and the Nutanix Cloud Availability Zone (AZ). Its primary role is to provide a secure, encrypted tunnel for replicating recovery points and managing disaster recovery orchestration metadata. When utilizing standard IPSec for this traffic, the system architecture imposes certain throughput limitations based on the processing capabilities of the gateway instance. For a single Nutanix Cloud Gateway, the maximum supported bandwidth for IPSec-encrypted traffic is 1 Gbps.

This 1 Gbps limit is an important consideration for BCDR planning, as it defines the maximum " Snapshot-on-wire " throughput available for a given protection policy. If an organization has a high data change rate (churn) that exceeds 1 Gbps of sustained replication traffic, the Recovery Point Objective (RPO) may be at risk of lag. While the physical network link might be 10 Gbps or higher, the encryption overhead of the IPSec protocol within the gateway VM is the limiting factor. To achieve higher aggregate throughput, administrators would need to deploy multiple gateways or consider high-performance direct connectivity options if available. Understanding this 1 Gbps threshold ensures that administrators can accurately size their replication windows and manage expectations for initial seed transfers and ongoing delta synchronization to the cloud AZ.

Nutanix Metro Availability is a continuous availability solution designed to survive a site failure with zero data loss (Zero RPO). Because it operates on a synchronous replication model, every write performed by a virtual machine must be acknowledged by both the local and the remote clusters before the write is confirmed to the application. This process introduces a " latency penalty " equal to the network Round Trip Time (RTT) between the sites.

To maintain acceptable application performance and ensure cluster stability, Nutanix enforces a strict latency limit of 5 ms RTT between the primary and secondary clusters. If the latency exceeds this 5 ms threshold, the time spent waiting for remote acknowledgments would cause severe performance degradation for high-I/O workloads, such as the SQL or Oracle databases common in financial institutions. Furthermore, high latency increases the risk of the synchronous link being " broken " during periods of high data churn. While the Witness VM (used for tie-breaking) can tolerate much higher latency (up to 200 ms), the data-path link must remain under 5 ms to satisfy the technical requirements of the Metro Availability feature and the performance expectations of Tier-1 business-critical applications.

Failing over to Nutanix Cloud Clusters (NC2) on AWS involves moving workloads from a local, on-premises networking environment into a cloud-native Virtual Private Cloud (VPC). While Nutanix Flow Virtual Networking (FVN) allows the VMs to keep their internal IP addresses using overlay networks, external visibility requires integration with AWS networking components.

Standard internal IP addresses in an overlay network are not natively routable from the public internet or from the AWS VPC outside of the NC2 cluster. To bridge this gap, Nutanix uses " Floating IPs " . A Floating IP is an AWS Elastic IP that is mapped to the internal IP of the guest VM. When an external user attempts to reach the website, the traffic hits the AWS public-facing Floating IP, which is then translated and routed through the NC2 bare-metal nodes to the specific guest VM. This mapping can be automated as part of the Recovery Plan or configured manually post-failover. Without this configuration, the VMs will remain " isolated " from external traffic, even if the application is running perfectly and internal cluster connectivity is healthy. This demonstrates the importance of cloud-native networking knowledge when extending on-premises BCDR plans into public cloud environments like NC2.

Nutanix Disaster Recovery (formerly Leap) utilizes a " policy-driven " approach where virtual machines are protected based on assigned " Categories " rather than individual VM names. This allows for automated protection: any VM tagged with a specific category (e.g., App:Web) is automatically picked up by the associated Protection Policy. When existing VMs are replicating fine but new ones are not, and there are no network or storage errors, the root cause is almost always an administrative oversight during provisioning.

If the separate team that deployed the VMs followed a checklist that didn ' t include category assignment, those VMs exist as " unprotected entities " in the eyes of Prism Central. The Protection Policy will ignore them because they do not match the entity selection criteria defined in the policy ' s configuration. To resolve this, the administrator must simply apply the correct category to the new VMs. Once tagged, the Cerebro service will identify the new members during its next scan and initiate the initial full seed replication to the recovery site. This scenario highlights the critical need for cross-team coordination and the use of automated deployment tools (like Nutanix Self-Service or Calm) to ensure that security and BCDR requirements are applied consistently during the VM lifecycle.

Nutanix Self-Service Restore (SSR) allows end-users to recover files from snapshots without needing assistance from a cluster administrator. This functionality is enabled by Nutanix Guest Tools (NGT), which mounts a snapshot disk to the guest VM. For SSR to function, the Nutanix Guest Agent (NGA) must be able to recognize and mount the file systems present on the snapshots.

For Linux-based virtual machines, Nutanix SSR supports standard, modern journaling file systems. EXT4 is the most widely supported and common file system used across modern Linux distributions like RHEL, CentOS, and Ubuntu, and it is natively supported by the Nutanix SSR agent. While other file systems like XFS or older EXT versions might be supported by the OS itself, EXT4 is the primary validated target for SSR operations. NTFS (Option B) is primarily a Windows file system, though it can be mounted on Linux, it is not the native standard for Linux SSR. Supporting the native file system of the guest OS is critical because it ensures that file permissions and ownership (UID/GID) are preserved during the recovery process. This compatibility allows users to seamlessly mount the snapshot, navigate to the required directory, and restore deleted or corrupted files directly to their production environment.

In a Nutanix legacy Protection Domain (PD) configuration, three primary elements must be in place for data protection to occur: the entities (VMs), the remote site (destination), and the schedule. Even if a VM has been added to a Protection Domain (negating Option A), the cluster will not take any action unless it is explicitly told " when " and " how often " to do so.

If no new snapshots are being created locally and replication is not starting, the most likely issue is that no " Schedule " has been configured for that Protection Domain. The schedule defines the RPO (e.g., hourly, daily) and the retention period for the snapshots. Without an active schedule, the Cerebro service simply has no instruction to initiate the snapshot process. While a missing vStore mapping (Option C) would cause the " replication " to fail, the " local snapshots " would still be created in the PD. Similarly, low bandwidth throttling (Option D) would cause the replication to move slowly or time out, but it would not prevent the initial creation of the local recovery points. Therefore, verifying the presence of a valid, active replication schedule is the mandatory first step for an administrator when a Protection Domain appears to be " idle " despite having VMs assigned to it.

Volume Groups (VGs) are used in Nutanix to provide block storage to virtual machines, often for high-performance applications like SQL Server. VGs can be attached to VMs in two ways: " In-guest " (using iSCSI initiators) or " Hypervisor-attached " (where the hypervisor handles the connection).

In a disaster recovery scenario managed by Recovery Plans, there is a technical distinction in how these attachments are handled. While Nutanix Disaster Recovery can replicate the data in the Volume Groups and recover the VM itself, VGs that are " hypervisor-attached " do not currently support automatic reattachment as part of the orchestration workflow in all AOS versions. When the failover completes, the VM powers on at the recovery site, but the VGs are not connected to the VM at the hypervisor level. This causes the application (SQL) to find its data disks missing, resulting in the database staying offline. To resolve this, the administrator must manually reattach the Volume Groups to the VM at the recovery site using the Prism interface or a script. This limitation highlights the need to verify specific application storage configurations during DR testing. For automated recovery of VGs, using " In-guest " iSCSI attachments is often preferred because the iSCSI initiator within the guest OS can re-establish the connection to the new cluster ' s Data Services IP once the network is active.

A multisite deployment in Nutanix Disaster Recovery allows a single primary cluster to replicate its workloads to multiple destination availability zones (AZs) for increased resilience. While a single " Protection Policy " can define the replication schedules to both recovery sites, the " Recovery Plan " is the separate object responsible for the actual orchestration of the failover.

A Recovery Plan is fundamentally a point-to-point object; it defines how to move workloads from Site A to Site B. If an administrator wants the ability to failover to either of two different recovery AZs, they must create two distinct Recovery Plans: one for AZ1 and another for AZ2. Each plan will contain the site-specific network mappings, boot sequences, and IP mappings relevant to that specific destination. If the administrator has only created a recovery plan for the first AZ, they will be able to see recovery points at both locations but will only have a " target " defined for the first site. The second AZ will not appear as a valid destination in the failover menu until its specific Recovery Plan is built and validated. This modular design ensures that each DR site can have its own customized networking and application orchestration logic.

Synchronous replication is a high-performance disaster recovery solution designed to provide a Zero Recovery Point Objective (RPO) by mirroring every write operation in real-time between two clusters. Because the system must wait for the remote site to acknowledge every write before confirming the operation to the application, the underlying storage media must be capable of extremely low-latency I/O.

On hybrid clusters that utilize both Solid State Drives (SSDs) and Hard Disk Drives (HDDs), Nutanix mandates a minimum of one SSD per node to support synchronous replication. The SSD tier is used to store metadata and provide the high-speed " Oplog " where incoming writes are initially landed before being drained to the HDD tier. If a cluster were to attempt synchronous replication using only HDDs (Option D), the mechanical latency of the spinning disks combined with the network Round Trip Time (RTT) would cause severe application performance degradation, failing the business requirement. Even though the cluster has large capacity requirements (100 TB), the presence of an SSD tier is a technical prerequisite for the software to even enable the synchronous consistency feature. Therefore, ensuring at least one SSD exists in every node of the hybrid cluster is the critical storage validation step required before implementation.

Transitioning from legacy Protection Domains (which are cluster-centric) to modern Prism Central-based Protection Policies (which are management-plane centric) is a common task during Nutanix environment upgrades. This migration involves a " handoff " where the responsibility for snapshots and replication shifts from the local Cerebro service on the cluster to the global orchestration provided by Prism Central.

The primary risk during this migration is the creation of a " protection gap. " If an administrator deletes the legacy snapshots immediately after assigning a new policy, and that new policy has not yet successfully completed its first replication cycle, the virtual machine is effectively unprotected. If a disaster occurs at that exact moment, there would be no valid recovery points at the destination site to restore from. Nutanix best practice dictates that legacy artifacts must be preserved until the new system is verified as operational. Once the first recovery point for the VM appears in the Prism Central " Recovery Points " tab, it confirms that the new Protection Policy has successfully captured the VM ' s state and replicated it to the recovery AZ. At this stage, the legacy snapshots in the Protection Domain become redundant and can be safely deleted to reclaim storage space without compromising the organization ' s ability to meet its Recovery Point Objective (RPO) and Recovery Time Objective (RTO).

VM-VM Anti-Affinity rules are used in Nutanix to ensure that specific virtual machines (such as the nodes of a database cluster or redundant web servers) never run on the same physical host, thereby protecting the application from a single host failure. These rules are typically managed through Prism Central (PC) in a modern deployment.

However, in the context of Disaster Recovery (DR) between two different clusters or availability zones, there is a technical limitation regarding the replication of these affinity policies. While the virtual machine ' s data and primary metadata (CPU, RAM, Disks) are replicated by the Cerebro service, the affinity rules are often considered local cluster constraints. In many versions of Nutanix Disaster Recovery (specifically the initial iterations of Leap/Prism Central-based DR), these affinity policies are not natively " synchronized " or enforced at the recovery site during a failover. This means that when the VMs are recovered at the destination cluster, the hypervisor ' s dynamic scheduler (ADS) may place them on the same host if that host has the most available resources. To maintain anti-affinity at the recovery site, an administrator would typically need to manually recreate the affinity policies at the destination AZ or use specialized orchestration scripts. Understanding this limitation is vital for ensuring that high-availability requirements are still met after a disaster recovery event has occurred.

This question evaluates the administrator ' s ability to map business RPO (Recovery Point Objective) requirements to the appropriate Nutanix technology. A requirement for " maximum 15 minutes of data loss " means the system must replicate data at least every 15 minutes. Nutanix offers two primary ways to achieve this: traditional Asynchronous replication and Nearsync replication. While Async (Option D) can technically be configured for a 15-minute interval, it relies on standard, full snapshots which can be heavy and resource-intensive, potentially leading to performance lag if the data churn is high. Nearsync Replication (Option B) is specifically designed for these low-RPO scenarios (from 20 seconds up to 15 minutes) by using " Lightweight Snapshots " (LWS). LWS minimize the storage and I/O overhead compared to traditional snapshots, making Nearsync the more reliable and robust choice for Tier-1 applications. Furthermore, Nearsync is built to handle the constant, high-frequency data transfers required to maintain a strict 15-minute window without impacting application performance. Therefore, Nearsync with a 15-minute RPO is the " best " configuration because it is the purpose-built solution for the stated performance and data protection goals.

By default, Nutanix clusters use the same management network interface (eth0) for administrative tasks and replication traffic. In high-traffic environments or over narrow WAN links, replication can consume all available bandwidth, impacting management accessibility or other VM traffic. When a new dedicated high-speed link (such as a 10G connection) is added specifically for BCDR, the cluster must be instructed to move its replication traffic to that new path.

The Nutanix solution for this is " Network Segmentation for DR. " This feature allows the administrator to define a new, dedicated network for disaster recovery and replication. By configuring segmentation, the administrator assigns a specific virtual interface (typically ntnx0) and a dedicated set of IP addresses for each Controller VM (CVM) on the new 10G link. Once enabled, the Cerebro and Stargate services will bind to these new " segmented " IPs for all site-to-site data transfers. This effectively offloads replication from the management network, ensuring that the 10G link is fully utilized for BCDR while the management network remains stable for administrative tasks. This configuration provides both better performance for replication (meeting aggressive RPOs) and better overall cluster security and stability through traffic isolation.

In modern Nutanix Disaster Recovery (using Prism Central-based Protection Policies), storage placement is often governed by Storage Policies rather than simple container-to-container mapping. Storage Policies use Nutanix Categories to determine where a VM should be placed to meet performance or compliance requirements. In this scenario, the VMs were successfully replicated and recovered, but they were placed in " all the storage containers " rather than the intended FIN-PROD container at the Marseille site. This indicates that while the data arrived, the " destination placement rules " were missing. If a storage policy was created on the Paris cluster to keep these VMs in FIN-PROD but that policy was not mirrored or assigned at the Marseille cluster, the system defaults to the Acropolis Dynamic Scheduler (ADS) placement. ADS will distribute the VMs across all available containers to balance load and capacity, leading to the " scattered " result observed. To resolve this and ensure consistent failover behavior, the administrator must ensure that both the categories and the corresponding storage policies are identical and active at both availability zones. While a missing container (Option C) would typically cause a replication failure rather than scattered recovery, a missing storage policy specifically explains the lack of orchestrated placement during the failover event.

Nutanix Recovery Plans allow for granular control over networking during different failover scenarios. The " Test Network " setting is specifically designed to provide an isolated sandbox for validation. If an administrator incorrectly maps this test network to the actual " Production Network " VLAN at the recovery site, they are effectively bypassing the safety mechanisms intended to isolate the test.

When the Test Failover is executed, the test VMs will boot up on the production VLAN using the exact same IP addresses they have at the source site. If the primary site is still online (which is the case during a test), the existence of two identical VMs with identical IPs on the same logical network will cause immediate IP address conflicts. This can disrupt live production traffic, cause routing instability, and potentially lead to service outages at both sites. The system does not " automatically " isolate traffic (Option B) if the administrator has explicitly directed the test VMs to the production segment. This scenario emphasizes the critical importance of network planning in BCDR: a " Test " network must always be an isolated, non-routable segment to ensure that validation remains non-disruptive as intended.

Nutanix Metro Availability is a continuous availability solution that provides a zero Recovery Point Objective (RPO) and near-zero Recovery Time Objective (RTO) across two sites. For this configuration, there are two distinct latency requirements that administrators must manage. First, the Round Trip Time (RTT) latency between the two storage clusters participating in the synchronous replication must be less than 5ms to ensure application performance is not significantly degraded by the write-synchronization process. Second, the Witness VM, which acts as an external arbitrator to prevent " split-brain " scenarios during a site failure, has a more flexible latency requirement. The maximum RTT between the Witness VM and the managed Protection Domains (PDs) must be less than 200ms. This higher threshold allows the Witness to be placed in a separate geographical region or a third-party cloud availability zone (AZ), ensuring it remains independent of the failure domain affecting the primary and secondary sites. If the latency to the Witness exceeds 200ms, the cluster may be unable to quickly acquire the lock necessary to automate failover, potentially causing the Metro Availability configuration to fail during an actual disaster event.

In a Nutanix Disaster Recovery context, choosing the right failover type is critical during a ransomware event. A " Planned Failover " is designed for maintenance scenarios. Its first step is to synchronize the current data state to the recovery site to ensure zero data loss. In a ransomware attack, the " current state " is encrypted or corrupted data. If an administrator performs a Planned Failover at 2:45 AM, the system will capture and replicate the encrypted files, effectively " backing up " the ransomware to the recovery site.

To successfully recover from ransomware, the administrator must perform an " Unplanned Failover. " Despite the name, this failover type is the correct choice for this scenario because it allows the administrator to manually select a historical recovery point from the destination cluster ' s storage. By choosing a recovery point from 1:00 AM (before the attack began), the administrator can restore the virtual machines to a known-good, unencrypted state. This is a primary use case for maintaining a multi-day or multi-week snapshot retention policy. While a " Test Failover " (Option A) could be used to verify the 1:00 AM snapshot in an isolated network first, the ultimate production recovery is achieved via an Unplanned Failover that specifically targets a point-in-time snapshot that predates the infection.

The Nutanix Cerebro service manages replication by scheduling snapshots and transferring the delta changes to the remote site. In an asynchronous replication schedule, if a new snapshot is taken before the previous one has finished replicating, the system faces a choice. To ensure that the most current data is available at the recovery site as quickly as possible, Nutanix may " skip " the replication of an older pending snapshot in favor of starting the replication of the newest one.

In this scenario, all infrastructure components (bandwidth, ports, hardware) are verified as healthy and have supported successful replication for months. This points toward a data-level issue rather than a connectivity issue. A " significant one-time surge in change rate " (Option D) means the VM generated a massive amount of new data blocks (churn) in a short period. Because the " Snapshot-on-wire " transfer size is directly tied to the number of changed blocks, this specific snapshot would take much longer than usual to replicate. If the transfer time exceeds the RPO interval, the next scheduled snapshot will trigger. The Cerebro service will see that it has a newer recovery point ready to go and will skip the " stuck " or slow-moving older snapshot to maintain the most recent RPO possible at the target. This behavior generates the " Replication was skipped " warning to notify the administrator that the historical sequence was interrupted due to a temporary imbalance between data churn and available throughput.

Replication between Nutanix clusters involves the transfer of both actual data blocks and the metadata required to manage the consistency and scheduling of those snapshots. Two specific services running on the Controller VMs (CVMs) are responsible for these tasks: Stargate and Cerebro.

Port TCP 2009 is used by the Stargate service. Stargate is the primary I/O manager for the Nutanix cluster, and in a BCDR context, it handles the " Snapshot-on-wire " data transfer. This is the port where the actual bits of the virtual disks are sent from the source cluster to the destination cluster. Port TCP 2020 is used by the Cerebro service. Cerebro is the replication engine that handles protection domain orchestration, snapshot scheduling, and cross-site handshake operations. Without port 2020, the clusters cannot synchronize their schedules or confirm that a recovery point has been successfully stored at the remote site. While other ports like 9440 (Prism management) and 22 (SSH) are used for administrative tasks, ports 2009 and 2020 are the mandatory " data path " and " control path " ports required for the continuous and successful replication of workloads between availability zones.

The Nutanix architecture relies on specialized services running on each Controller VM (CVM) to handle specific cluster operations. The " Cerebro " service is the primary engine responsible for data protection and replication tasks across the Nutanix cluster. It manages recovery points, snapshot schedules, and the actual transfer of data between the local cluster and remote sites. Consequently, the cerebro.INFO log file is the most vital resource for an administrator troubleshooting replication health or site-to-site connectivity. This log records critical events such as the start and completion of replication jobs, handshake failures between clusters, and timeout errors during snapshot transfers. If an administrator notices that replication is " hung " or failing consistently, cerebro.INFO provides the granular detail needed to see if the issue is a network timeout, a remote service unavailability, or a metadata synchronization error. Reviewing this log helps ensure that the connection stability is maintained and that any deviations in replication performance are identified before they impact the Recovery Point Objective (RPO). Other logs like genesis.out (Option B) are more focused on service startup and configuration rather than the ongoing health of the BCDR replication stream.