Explanation:
IGMP (Internet Group Management Protocol) Snooping is a network feature designed specifically to optimize multicast traffic at the Layer 2 (switch) level.

Here's how it works in the context of AHV:

Without IGMP Snooping:
A virtual switch would flood all multicast traffic to every port (and thus to every VM) in the broadcast domain, regardless of whether the VMs want to receive it. This wastes network bandwidth and host CPU cycles.

With IGMP Snooping:
The virtual switch "snoops" on IGMP messages exchanged between VMs and the network router. It listens for "IGMP Join" and "IGMP Leave" messages from the VMs.

The Optimization:
By listening to these messages, the virtual switch learns which VMs are members of which multicast groups. It then intelligently forwards multicast traffic only to the ports where interested VMs are connected, instead of flooding it to all ports.
This directly fulfills the requirement to "only forward multicast traffic to the VMs that need to receive it."

Why the Other Options Are Incorrect

A. LACP (Link Aggregation Control Protocol):
This is a bonding protocol used to aggregate multiple physical network links for increased bandwidth and redundancy. It operates at a lower level and has no intelligence for managing or optimizing multicast traffic flows.

B. UDP (User Datagram Protocol):
This is a core transport layer protocol (Layer 4). Multicast often uses UDP, but the protocol itself is not a feature that can be configured on AHV to control traffic forwarding. It is simply the type of traffic being sent.

D. Network Segmentation:
While segmenting the network (e.g., using VLANs) can contain broadcast/multicast traffic to a specific segment, it is a blunt instrument. Within that segment, multicast traffic would still be flooded to all VMs. It does not provide the fine-grained, dynamic optimization that IGMP Snooping does.

Reference:
The Nutanix AHV Networking Guide discusses virtual switch configuration and features. Enabling IGMP Snooping on the AHV host bridge (the virtual switch) is the standard and recommended method for efficiently handling IP multicast traffic, preventing unnecessary flooding and conserving network resources within the cluster.

Explanation:
In a Nutanix cluster configured with Replication Factor 2 (RF2), the system maintains two copies of each data block across different nodes to ensure fault tolerance. While RF2 clusters can operate with a minimum of 3 nodes, you need at least 4 nodes to safely remove one without violating RF2 requirements.

✅ Why Option C (4 nodes) Is Correct
RF2 requires two distinct nodes to store each data block.
A 3-node RF2 cluster is the minimum viable configuration, but removing one node would leave only 2, which breaks RF2 redundancy.
With 4 nodes, removing one still leaves 3 nodes—enough to maintain RF2 and re-replicate data safely.
Nutanix prevents node removal if doing so would compromise data protection or cluster health.

📚 Reference:

Nutanix Cluster Operations Guide:
“To remove a node from an RF2 cluster, a minimum of four nodes is required to maintain redundancy.”
Nutanix Support KB – Node Removal Requirements

❌ Why Other Options Are Incorrect

A. 2
RF2 cannot operate with only 2 nodes because it needs two separate nodes for each data block.
A 2-node cluster would violate RF2’s fault tolerance model and is not supported.

B. 3
Although RF2 clusters can run with 3 nodes, removing one would leave only 2.
This breaks the replication model and risks data loss.
Nutanix blocks node removal in this scenario.

D. 5
While valid, 5 nodes exceed the minimum requirement.
You can remove a node from a 5-node RF2 cluster, but it’s not the minimum needed.
The question asks for the lowest number that still allows safe removal.

Summary:
To safely remove a node from an RF2 Nutanix cluster, you must have at least 4 nodes. This ensures that after removal, the remaining nodes can still maintain RF2’s two-copy redundancy. Options A and B fail to meet RF2 requirements post-removal, and option D is valid but not minimal. Nutanix enforces this to protect cluster integrity and prevent data loss.

Explanation:
Intelligent Operations is a feature within Prism Central that provides advanced monitoring, alerting, and reporting. A core component of this feature is the ability to generate and schedule reports.

Weekly Performance Summary:
The "Infrastructure Performance" report is a specific type of Intelligent Operations report that provides a high-level summary of cluster health, capacity, and performance over a selected period.

Scheduling:
This feature allows an administrator to configure the report to be generated and sent via email on a recurring schedule, such as weekly. This delivers the exact "weekly message about infrastructure performance summary" described in the question.

Why the Other Options Are Incorrect

A. Admin Center Life Cycle Manager:
This is incorrect. Life Cycle Manager (LCM) is used for planning and performing firmware and software upgrades. It does not generate or send scheduled performance summary reports.

B. Prism Central Syslog:
This is incorrect. Syslog is a protocol for forwarding log and event messages to a central syslog server for analysis and storage. It provides a real-time or near-real-time stream of individual events, not a consolidated, human-readable weekly summary report.

C. Infrastructure VMs List:
This is incorrect. While you can view a list of Infrastructure VMs (like CVMs, Prism Central, etc.) in Prism, this is a static inventory view. It is not a scheduled reporting feature that summarizes performance metrics.

Reference:
This functionality is documented in the Nutanix Prism Central Guide under the "Intelligent Operations" section. The guide details how to create, customize, and schedule various reports, including the Infrastructure Performance report, to be automatically delivered to specified email addresses on a daily, weekly, or monthly basis.

Explanation:
Storage Policies in Nutanix are a core feature of the Acropolis Distributed Storage Fabric (ADSF). They allow an administrator to define a set of storage characteristics (attributes) that can be applied to individual virtual disks (vDisks) on a per-VM basis.

The primary storage attributes managed by Storage Policies are:

Replication Factor (RF):
This defines the number of copies of data maintained across the cluster (e.g., RF2 for two copies, RF3 for three copies). This is the most common attribute set by a storage policy, allowing different levels of data redundancy for different VMs.

Encryption:
Storage Policies can specify whether a vDisk should be encrypted at rest. This allows for granular control, where only sensitive VMs have their storage encrypted, as opposed to enabling cluster-wide encryption.
By using Storage Policies, administrators can provide different levels of service (e.g., Gold, Silver, Bronze) for different applications without needing to create separate storage containers.

Why the Other Options Are Incorrect
A. Storage Containers and Volume Groups:
This is incorrect. Storage Containers are managed independently and act as a logical pool of storage with their own settings (like compression, erasure coding). Storage Policies are applied within a container to vDisks. Volume Groups (VGs) are a separate entity for attaching external storage to VMs and are not managed by the Storage Policy framework used for native ADSF vDisks.

C. Shares and Object Stores:
This is incorrect. "Shares" typically refer to SMB or NFS file shares, which are not managed by the VM-centric Storage Policy system. Object Stores (e.g., S3-compatible buckets) are a different service (Objects) and are configured and managed separately from the block storage governed by Storage Policies.

D. Data Protection and Security:
This is too vague and partially incorrect. While Encryption (a security feature) is an attribute, "Data Protection" in Nutanix generally refers to features like snapshots, Protection Domains, and Async/Metro DR. These are configured at the VM or group level, not as an attribute of a storage policy applied to an individual vDisk.

Reference:

The Nutanix Acropolis Storage documentation explicitly defines Storage Policies as the mechanism to control the Resilience Factor (RF) and Encryption settings for individual vDisks. This allows for granular data redundancy and security settings without the need to place VMs in different storage containers.

Explanation:
In a Nutanix hybrid (SSD + HDD) cluster, the SSD tier serves a critical performance function:

Caching and Tiering:
The SSDs act as a cache (Curator) and a performance tier for "hot" data and metadata. All write I/O and the most frequently read data (hot data) are served from the SSD tier.

Impact of High Utilization:
When SSD utilization is consistently above 75%, it indicates that the cache is becoming full. This leads to:
Reduced Caching Efficiency:
There is less free space available to cache new write I/O and promote hot read data from the HDD tier.
Increased HDD Access:
More I/O operations will be forced to go to the slower HDD tier to read data that can no longer fit in the SSD cache.

Higher Latency:
Since HDDs have significantly higher latency than SSDs, this shift of I/O from SSD to HDD results in a noticeable increase in the average I/O latency for VMs and applications.
This performance degradation is the direct and most common impact of the described alert.

Why the Other Options Are Incorrect

A. The cluster is unable to sustain an SSD disk failure.
This is incorrect. The ability to sustain a disk failure is determined by the Resiliency Factor (RF) and the state of the data rebuild process. While high utilization can slow down rebuild times, it does not inherently break the cluster's fault tolerance. A cluster with high SSD utilization can still survive a disk failure if it has sufficient free capacity elsewhere to maintain RF.

B. The cluster may be nearly out of storage for metadata.
This is incorrect and too specific. Metadata is stored on the SSD tier, but high overall SSD utilization doesn't exclusively target metadata. The alert is for general "SSD utilization," which includes user data, cache data, and metadata. The primary risk is general performance degradation, not a specific metadata exhaustion.

C. The cluster is at risk of entering a read-only state.
This is incorrect. A cluster enters a read-only state when it is critically low on total storage space and cannot safely process writes without risking data loss. The alert is specifically about SSD utilization in a hybrid cluster. The HDD tier likely still has free capacity for storing persistent data. The cluster as a whole is not out of space; rather, it is running low on its high-performance resources.

Reference:

This behavior is documented in the Nutanix Acropolis Storage documentation regarding hybrid cluster architecture and performance tuning. The guides explain that the SSD tier is critical for low-latency I/O and that monitoring its utilization is key to maintaining performance. Consistently high SSD utilization is a leading indicator of potential latency issues as I/O spills over to the slower HDD tier.

Explanation:
When creating a VM template in Nutanix Prism, the guest customization feature provides several options to configure the operating system of a VM upon its first boot from that template. These options are:

Sysprep: Used primarily for customizing Windows VMs. It allows for tasks like setting the hostname, joining a domain, and running specialized commands to generalize the Windows instance.

Cloud-init:
A widely supported industry standard for Linux and modern Windows Server versions. It uses a user-data script (often in YAML format) to handle network configuration, user creation, package installation, and running custom scripts.

Custom Script:
Allows you to directly input a script (e.g., a shell script for Linux or a PowerShell script for Windows) that will be executed on the first boot.

Guided Script:
A user-friendly interface that provides form fields for common customization tasks (like setting the hostname, DNS, and domain), which Prism then converts into the appropriate script (Sysprep or cloud-init) in the background.
These options ensure that VMs deployed from a template are not identical clones but can be dynamically configured for their specific role and environment.

Why the Other Options Are Incorrect

B. Bash, Powershell:
These are specific scripting languages, not the available options within the Prism template customization interface. A Bash script would be used within the "Custom Script" or "Cloud-init" option for Linux, and a PowerShell script would be used within the "Custom Script" or "Sysprep" option for Windows.

C. Python, YAML:
Similar to option B, these are technologies used to implement customization, not the direct options in Prism. Python scripts can be run via a "Custom Script," and YAML is the common format for "Cloud-init" configuration files.

D. None, guest customization is not supported in Nutanix templates.
This is factually incorrect. Guest customization is a core and well-supported feature when creating VM templates on the Nutanix AHV platform.

Reference:

The availability of these specific guest customization options (Sysprep, Cloud-init, Custom Script, Guided Script) is documented in the Nutanix Prism Web Console Guide and the AHV VM Management guide. This functionality is accessible during the VM template creation workflow in Prism Element and Prism Central.

Explanation:

✅ Why B is Correct

Nutanix Categories are the most efficient and scalable way to organize and manage workloads across business units and application tiers in Disaster Recovery (DR). Categories allow administrators to tag VMs based on logical attributes such as business unit, application tier, environment (e.g., prod/dev), or SLA requirements. These tags can then be used to dynamically group VMs into Recovery Plans in Nutanix Leap.
Categories support multi-tenancy and policy-driven DR orchestration.
They eliminate the need for manual VM selection or static naming conventions.
Recovery Plans can be built using category filters, ensuring automated inclusion of relevant VMs as environments evolve.

📚 Reference:

Nutanix Leap DR Guide: “Categories provide a scalable way to organize VMs for recovery planning across business units and applications.”
Nutanix Documentation – Using Categories in Recovery Plans

❌ Why Other Options Are Incorrect

A. Utilize a VM naming schema that allows sorting
Naming conventions are helpful but not scalable or dynamic.
They rely on manual consistency and do not support policy-based automation.
Categories are more flexible and robust for DR orchestration.

C. Utilize a 1:10 ratio of Recovery Plan to VMs
There is no recommended fixed ratio of Recovery Plans to VMs.
Recovery Plans should be based on logical grouping and SLA, not arbitrary ratios.

D. Utilize RESTful APIs to script creation of Recovery Plans
While APIs are powerful for automation, they do not solve workload organization.
Categories can be used in conjunction with APIs, but they are the core method for organizing VMs.

Summary

For multi-tier applications spanning multiple business units, Categories offer the most efficient, scalable, and policy-driven way to organize VMs in Nutanix Disaster Recovery. They enable dynamic grouping, simplify Recovery Plan creation, and support enterprise-grade DR strategies. Other options either lack automation, are static, or do not address workload organization directly.

Explanation:
The automatic discovery of new nodes during a cluster expansion relies on network communication protocols. The specific requirement is:

Discovery Mechanism:
New, unconfigured nodes (often referred to as "Phoenix" nodes) use IPv4 multicast messages to announce their presence on the network.

Cluster Communication:
Existing nodes in the target cluster listen for these multicast announcements. When an existing node detects a new node's multicast message, it facilitates the process of adding the new node to the cluster.

Network Prerequisite:
For this communication to work, the physical network switches connecting the new nodes to the existing cluster must be configured to allow IPv4 multicast traffic to flow between the segments hosting the existing cluster and the new nodes. If multicast is blocked by the network infrastructure, the automatic discovery process will fail.

Why the Other Options Are Incorrect

A. New nodes must have the same hypervisor version.
This is incorrect. While it is a best practice for operational consistency, it is not a strict requirement for the initial discovery and joining process. Life Cycle Manager (LCM) can be used after the node is joined to standardize the hypervisor version.

B. IPv6 multicast must be allowed on physical switches.
This is incorrect. The Nutanix cluster formation and discovery process primarily uses IPv4 multicast, not IPv6.

C. New nodes must have the same AOS version.
This is incorrect and often false. In fact, one of the key features of Nutanix expansion is that you can add nodes running a newer compatible AOS version to an existing cluster. The cluster's LCM will then use a "rolling upgrade" strategy to update the existing nodes. The new node's AOS version must be compatible, but not necessarily identical.

Reference

The Nutanix Cluster Management guide explicitly lists network requirements for expanding a cluster. It specifies that the physical network must support and forward IPv4 multicast traffic (e.g., IGMP) for the automatic node discovery process to function correctly. Blocking multicast is a common cause of expansion failures.