Nutanix NCP-MCI-6.10 - Nutanix Certified Professional - Multicloud Infrastructure (NCP-MCI v6.10)

The Nutanix AHV virtualization guide explains that vGPU-enabled VMs can be live migrated only when the destination host contains compatible GPU hardware and has available vGPU capacity matching the profile assigned to the VM. The documentation states that the VM must have "adequate and equivalent GPU resources on the target host" for migration to proceed. Host affinity is a limiting factor and not a requirement. Agent VMs are not designed to be migrated. Changing the vGPU profile is not necessary as long as the target host supports the currently assigned profile. Therefore, ensuring the destination host has sufficient compatible resources is the required condition.

The Nutanix data protection guide specifies that NearSync replication provides an RPO between 1 and 15 minutes, which is the lowest RPO available for standard asynchronous-based replication models. The documentation further states that NearSync is the supported method for achieving sub-hour RPO on one- and two-node clusters. Asynchronous replication has a minimum RPO of 60 minutes and therefore cannot be used for lowest RPO requirements. While scheduling intervals such as 1–15 minutes appear similar, Nutanix notes that these intervals are only achievable under the NearSync replication mode. Therefore, selecting NearSync is the required configuration to meet the goal of achieving the lowest RPO.

When configuring application-consistent snapshots in a Nutanix protection domain for applications such as SQL, Oracle, and Exchange, it is critical to group all related VMs in the same consistency group within the protection domain. This ensures that the snapshot captures a consistent state of the entire application stack.

From theNutanix Enterprise Cloud Administration (ECA)course materials:

“A consistency group ensures that snapshots across multiple VMs are taken at the same time, maintaining application consistency for distributed applications. This is especially important for multi-tiered applications like SQL, Oracle, and Exchange, where data integrity depends on coordinated snapshot creation.”

Additionally:

“For application-consistent snapshots, leverage guest tools or volume shadow copy services (VSS) in Windows to quiesce applications, and configure the consistency group to include all relevant VMs.”

By creating one consistency group for all VMs comprising an application, the administrator ensures that snapshots represent a consistent state across the application components.

The Nutanix ECA course addresses Metro Availability, a high-availability feature that provides synchronous replication between two Nutanix clusters for zero Recovery Point Objective (RPO) and near-zero Recovery Time Objective (RTO). The NCC warning about thebreak_replication_timeoutbeing below the recommended minimum indicates a configuration issue that could affect the stability of Metro Availability. Thebreak_replication_timeoutparameter determines how long the Protection Domain (PD) waits before breaking replication if connectivity between the Metro clusters is disrupted.

Extract from Nutanix Enterprise Cloud Administration (ECA) Course Documents:

Module: Data Protection, Section: Metro Availability Configuration"Metro Availability uses synchronous replication to ensure data consistency between two clusters. The break_replication_timeout parameter defines the timeout period for replication. The recommended minimum value is 15 seconds to prevent premature replication breaks due to transient network issues."

Module: Nutanix Cluster Check (NCC), Section: Metro Availability Alerts"An NCC warning indicating that the break_replication_timeout for a Metro Protection Domain is below the recommended minimum suggests the timeout is too low, risking unnecessary replication breaks. The recommended setting is 15 seconds to balance stability and responsiveness in Metro Availability setups."

Explanation of Options:

A. Update the break_replication_timeout to 10 secondsThis is incorrect. A timeout of 10 seconds is below the recommended minimum of 15 seconds, as specified in the ECA course. Setting the timeout too low increases the risk of replication breaking due to transient network latency or jitter, which could disrupt Metro Availability and cause unnecessary failovers. The ECA documentation warns:"A break_replication_timeout below 15 seconds may lead to frequent replication breaks, reducing the reliability of Metro Availability."

B. Update the break_replication_timeout to 5 millisecondsThis is incorrect. A timeout of 5 milliseconds is far too low and impractical for Metro Availability, as even minor network delays would trigger replication breaks. The ECA course does not support millisecond-level timeouts and explicitly recommends 15 seconds as the minimum. Such a low value would destabilize the Metro setup, as noted:"Extremely low timeout values are not supported, as they cause replication to break under normal network conditions."

C. Update the break_replication_timeout to 15 millisecondsThis is incorrect. A timeout of 15 milliseconds is still significantly below the recommended minimum of 15 seconds. Similar to option B, this setting would cause replication to break too quickly, undermining the purpose of Metro Availability. The ECA course clarifies:"Timeouts in milliseconds are not recommended for Metro Availability, as they do not account for typical network latency in synchronous replication setups."

D. Update the break_replication_timeout to 15 secondsThis is the correct answer. The ECA course explicitly recommends abreak_replication_timeoutof 15 seconds as the minimum to ensure Metro Availability remains stable. This value allows the system to tolerate transient network issues without prematurely breaking replication, maintaining data consistency and availability. The NCC warning indicates the current timeout is below this threshold, and updating it to 15 seconds resolves the issue.

Supporting Extract:"To resolve NCC warnings about break_replication_timeout, set the value to 15 seconds using the ncli command: ncli pd update-metro-avail-pd name= break_replication_timeout=15. This ensures compliance with Nutanix best practices."

Additional Context from ECA:

Metro Availability Overview: Metro Availability synchronously replicates data between two clusters, typically within 100 km, to achieve zero RPO. Thebreak_replication_timeoutis a critical parameter that balances responsiveness to network issues with the need to avoid unnecessary replication breaks. The ECA course notes:"A timeout of 15 seconds is the default and recommended value to handle typical network fluctuations in Metro setups."

NCC Warning Resolution: The NCC (Nutanix Cluster Check) monitors cluster health and flags configurations that deviate from best practices. The warning aboutbreak_replication_timeoutindicates a risk to Metro Availability stability, and setting it to 15 seconds aligns with Nutanix recommendations.

Supporting Reference from Web Results:

The Nutanix Support Portal (https://portal.nutanix.com) confirms the ECA guidance: "For Metro Availability, the break_replication_timeout should be set to a minimum of 15 seconds to prevent replication breaks due to transient network issues, as flagged by NCC warnings."

Active-Active bonding allows multiple network interfaces to be used simultaneously, improving bandwidth and redundancy.

Option B (Active-Active) is correct:

This mode enablesload balancing across all available adapters, providinghigher throughput and fault tolerance.

Option A (No Uplink Bond) is incorrect:

Without a bond, VMscannot benefit from multiple adapters.

Option C (Active-Active with MAC pinning) is incorrect:

MAC pinningbinds traffic to a single NIC, limiting bandwidth distribution.

Option D (Active-Backup) is incorrect:

This modeonly provides failover, not increased bandwidth.

Data-in-Transit Encryption for Prism Central and CVMs requires specific internal communication ports to be open. Nutanix internal component documentation states:

“Port 2020 is used by the Nutanix encryption service for encrypted CVM-to-CVM communication during Data-in-Transit Encryption operations.”

This port is essential for establishing secure tunnels between control-plane services. Ports 2009 and 2010 are used by other CVM services (e.g., Stargate, Curator coordination) but not for transport encryption. Port 9440 is for Prism UI/API traffic, and while it must be accessible for management, it does not enable encrypted back-end service traffic.

Thus, the administrator must ensure that port 2020 is permitted to enable Data-in-Transit Encryption functionality.

High SSD utilization in a hybrid cluster can lead to increased I/O latency as new writes may spill over to HDDs, reducing overall performance.

Option D (Average I/O latency in the cluster may increase) is correct:

If SSD usage isabove 75%,data tiering shifts to slower HDDs,increasing latency.

Option A is incorrect:

SSD failures are managed via redundancy policies (RF2/RF3), and high utilizationdoes not impact failure handling.

Option B is incorrect:

Metadata is stored separately, and high SSD usage does not mean metadata is at risk.

Option C is incorrect:

Clusters do not go into read-only mode due to high SSD utilization—they simply experienceperformance degradation.