NCP-MCI-7.5 Sample Questions Answers

Nutanix documentation describes Multicloud Snapshot Technology (MST/NMST) as a way to replicate native Nutanix snapshots directly to cloud object storage such as AWS S3 or Azure Blob, and specifically positions it as a cost-effective disaster recovery and data mobility approach. Nutanix best-practice content also says MST is especially suitable when an organization does not want to maintain a full secondary DR site and is looking for a low-cost recovery-point location in the cloud. That aligns exactly with the requirement in this question. ( portal.nutanix.com , portal.nutanix.com , portal.nutanix.com )

Adding an NC2 recovery location or full DRaaS design can be valid but usually carries more infrastructure cost than simply replicating snapshots to object storage. A custom runbook to clone recovery points to S3 is not the native Nutanix DR method being tested. So for an external cloud-based recovery point location with cost efficiency as the deciding factor, A is the authentic answer.

A Nutanix vDisk live migration moves the VM’s virtual disks into a new storage container while the VM remains online. For that to complete, the destination container must have enough available capacity to receive the migrated vDisk objects. If the target container lacks space, the migration can stall or fail because Nutanix cannot fully place the destination copy and complete the handoff. This makes insufficient storage space in the target storage container the most direct explanation. Nutanix migration and storage workflows consistently rely on adequate target-side capacity before the move can finish successfully. ( portal.nutanix.com )

The distractors are less precise. Source-side snapshots can be part of protection or internal workflow behavior, but the question is specifically about moving to a new container, so target capacity is the key gating factor. “Local storage container” is too vague and is not the normal object referenced in this workflow. Operationally, the right next step is to confirm free space, reservations, and any capacity constraints on the destination container before retrying the migration. That is why B is the best answer.

A required RPO of 5 minutes immediately points to NearSync replication, because Nutanix documents NearSync protection policies as supporting recovery points in 1–15 minute intervals. The second requirement—automatically protect new mission-critical VMs—points to a Protection Policy approach in Prism Central, because protection policies can use categories and policy-based assignment to scale protection dynamically as new matching VMs are added. Combining those two requirements makes Protection Policy with NearSync replication the correct answer. ( Nutanix )

Protection Domains can support NearSync in older or alternate workflows, but the phrase “new mission critical VMs should be automatically protected” is the clue that pushes this question toward policy-based protection rather than manually maintained membership. Asynchronous replication also fails the first requirement because its RPO target is not the 5-minute class represented by NearSync. Nutanix’s modern DR design encourages policy-driven protection for scalable, category-based administration. Therefore B is the best fit for both the RPO and automatic-inclusion requirements. ( Nutanix )

This question is about CPU feature compatibility, not storage or networking uniformity. Nutanix documentation and KB guidance explain that VM migration can fail when clusters contain nodes with different CPU generations or different exposed CPU features. A VM that is already running with features negotiated on a newer-generation CPU may not live-migrate successfully to an older or feature-different node until it is power-cycled or configured with advanced processor compatibility. That makes C the correct answer. ( portal.nutanix.com , portal.nutanix.com , portal.nutanix.com )

The important subtlety is that “same core count, memory, storage, subnet, and software version” does not guarantee CPU-feature equivalency. Firmware or ADS activity is not the primary issue described. Affinity rules can restrict placement, but the classic symptom after mixing repurposed nodes is CPU compatibility. Nutanix also documents that advanced processor compatibility settings persist across migrations and power cycles, reinforcing that CPU exposure is the real gating factor here.

Nutanix DR documentation for test failover states that recovered test VMs are brought up in a test virtual network at the recovery AZ, separate from production. Nutanix also documents that when multiple recovery directions or multiple recovery AZs are involved, recovery planning should be created per target direction and the required production and test network mappings need to exist for the recovery plan being tested. If recovery must be tested in both directions between the two clusters, each side effectively needs a production network and a test network. That makes two production networks and two test networks the most complete and correct answer.

This question is really about symmetry. One-direction failover testing can be done with a simpler mapping set, but the moment the requirement says test recovery in both directions, each cluster must be able to act as a recovery site and provide an isolated test-recovery network. That is why a single shared test network is not the best answer here. Nutanix’s recovery-plan structure and test-failover guidance support this interpretation. Therefore C is the correct and most practical answer.

When the requirement is to deploy multiple remote clusters efficiently from a central point, Nutanix’s intended tool is Foundation Central. Nutanix documentation describes Foundation Central as a way to manage several Foundation instances from a single pane of glass, allowing administrators to create clusters of remote nodes. That design matches the scenario perfectly: five newly purchased clusters, remote deployment, and a need for efficiency and centralized control. ( Nutanix Portal )

The other tools are more limited in scope. Foundation for Windows/Mac and a standalone Foundation VM are useful for imaging and cluster creation, but they do not provide the same centralized, multi-cluster orchestration model that Foundation Central provides. Phoenix ISO is a recovery or re-imaging tool and is not the normal answer for greenfield mass deployment. In short, this question is less about “which tool can build a cluster” and more about “which tool is purpose-built for remote, scaled deployment across many clusters.” Nutanix positions Foundation Central exactly for that use case, so B is the best answer. ( Nutanix Portal )

Nutanix supports storage policy–based encryption in Prism Central, which is exactly the right mechanism when only a subset of workloads needs encryption. Nutanix documentation explains that in addition to cluster-level encryption, administrators can configure storage policies to encrypt data at rest and apply those policies selectively. Since the PCI-DSS workloads are already logically groupable by category, the proper design is to create a storage policy with encryption enabled and then associate it with the category assigned to those VMs. That makes C correct. ( portal.nutanix.com , portal.nutanix.com , portal.nutanix.com )

Cluster-level encryption would encrypt too much, which violates the requirement to leave general-purpose VMs unencrypted. Guest OS BitLocker is a separate operating-system control, not the Nutanix platform answer being tested here. Container-centric thinking is older and less flexible than policy-based encryption. Nutanix’s policy model is the authentic platform-aligned solution for category-scoped encryption using an external KMS.

For AHV, when enabling LACP, Nutanix documentation specifies the Active-Active bond type, also described as balance-tcp. Nutanix explicitly ties LACP and link aggregation behavior to the Active-Active configuration and notes that this configuration sets LACP fallback to Active-Backup on AHV hosts. That makes C the correct bond type for an LACP-enabled design. ( Nutanix Portal )

It is also important to understand why Active-Active with MAC Pinning is wrong here. Nutanix documentation states that the MAC-pinning / balance-slb style must not be used with link aggregation protocols such as LACP. So even though both options sound like they distribute traffic, only one is compatible with LACP. Active-Backup provides redundancy, but not the LACP behavior being requested. “No Uplink Bond” obviously does not meet the requirement. Even without the missing exhibit from the uploaded file, the Nutanix networking documentation makes the intended answer clear: Active-Active is the bond type used when enabling LACP on AHV. ( Nutanix Portal )

Nutanix documentation for VM revert behavior in Prism Central states that categories are preserved by default when a VM is reverted to a recovery point. Nutanix further notes that if an administrator wants categories to revert to the state captured in the recovery point, that must be done by using a V4 API override rather than the standard default behavior. That directly rules out choices A and B in the ordinary UI-based revert scenario. The expected default behavior is preservation of current categories, not rollback of categories to historical values.

The same recovery-point workflow also preserves present-day management context such as ownership rather than reconstructing all metadata exactly as it existed at the earlier capture point. The reason is practical: reverting the VM is primarily about workload state recovery, not about undoing every later governance assignment unless explicitly requested through advanced controls. From an administrator’s point of view, this is safer because access-control and ownership context is not unexpectedly rolled backward during a restore operation. Therefore the best answer is C: both the VM’s categories and ownership are preserved at their current state.

This workload stores MP4 video files, which are typically large, sequential, and already compressed. For data with those characteristics, additional compression work often provides little benefit and can create unnecessary processing overhead. The best optimization is to disable compression for that workload through a storage policy so cluster resources are not wasted trying to compress data that is unlikely to shrink meaningfully. ( Nutanix Portal )

The other choices are weaker for this scenario. Capacity deduplication is generally most beneficial for repeated blocks or common-image patterns, not unique media streams. Erasure Coding is a capacity-efficiency feature, but on a small 3-node cluster and for a write-heavy recording pattern, it is not the primary performance optimization answer. Moving the data to Files changes the access method, but it does not address the specific inefficiency already identified in the storage settings. Nutanix exams often test whether the administrator can match the data service to the workload profile. Since MP4 content is already compressed, removing unnecessary compression overhead is the most targeted and sensible answer, making D correct. ( Nutanix Portal )

Nutanix documentation for Performing Failover in PD-based DR includes the exact error message quoted in the question: “Protection domain is already active on source site ... Please trigger migrate from source site to activate the protection domain.” Nutanix shows this message when the protection domain is still active on the source cluster and the proper action is a planned migrate from the source, not an activate at the target. In other words, the source cluster is still reachable and authoritative, so trying to activate the protection domain on the other side is the wrong operation. That makes D the correct answer. ( Nutanix Portal )

This is a classic Nutanix DR distinction between planned migration and unplanned failover/activation. If the primary cluster is still available, Nutanix expects the administrator to use migrate from the source site. Activation at the destination is reserved for failover conditions when the source cannot safely coordinate the move. The error itself tells you the reason if you read it carefully: the protection domain is still active on the source site. Therefore the root cause is that the administrator is trying to activate the PD while the primary cluster is still reachable and still owns the active copy, which is exactly option D. ( Nutanix Portal )

Nutanix Flow Network Security provides built-in Quarantine category values, including Strict and Forensic. Nutanix documentation states that Strict blocks all inbound and outbound traffic, while Forensic is specifically intended for situations where you want to quarantine the VM but still allow connectivity for forensic access. The documentation also notes that you can configure the allow behavior for forensic mode, which is exactly what this incident-response scenario requires. Therefore A is the correct answer.

This is a strong example of why built-in categories matter operationally. The organization wants a change that can be applied quickly during an incident and does not require redesigning existing policies. Assigning the VM to the Quarantine: Forensic category accomplishes both goals: it uses an existing system-defined control and preserves investigator access. Strict mode would over-isolate the VM and block the forensic tooling too. Thus the most authentic Nutanix response is A.

The error message is tied to resilience requirements. Nutanix documentation states that Replication Factor 3 requires at least five nodes so data remains online even if two nodes fail concurrently. If removing a node would drop the cluster below the minimum usable-node requirement, Nutanix blocks the operation. The safest and supported way to proceed during a hardware refresh is to add the replacement node first, restore compliant cluster capacity, and then remove the original node. ( Nutanix Portal )

The wrong answers all involve bypassing or weakening resilience inappropriately. Changing cluster or container RF to RF2 just to remove a node would reduce fault tolerance and is not the right answer for a SQL environment where resilience matters. Force-removal is also not the supported operational answer in a healthy refresh workflow. This question tests whether the administrator understands that node lifecycle actions are constrained by Nutanix fault-tolerance rules. When the cluster says it needs at least five usable nodes, the correct action is not to override the rule but to satisfy it by adding capacity first. Therefore A is the correct answer. ( Nutanix Portal )

Nutanix documentation explains that Projects in Prism Central provide logical groupings of infrastructure resources and user access, and that you can add users and user groups to a project while also allowing clusters and subnets in that project. That is exactly the mechanism used when a user group needs controlled access to resources drawn from multiple clusters. The project becomes the governance boundary that ties users, roles, and infrastructure scope together. Therefore A. Configure Nutanix Projects is the correct answer. ( Nutanix Portal )

The other choices are unrelated to access scoping across clusters. Advanced Network Controller deals with networking services and policy capabilities, not user resource assignment. Micro Services Platform is not the RBAC/governance feature for multi-cluster access. Availability Zone is relevant for placement and DR topologies, not for delegating cluster resources to user groups. Nutanix’s governance guidance is clear that Projects are the construct used to organize infrastructure resources and assign them to users or groups with the proper RBAC controls. So for a user group that needs access to resources from multiple clusters, the authentic Prism Central action is A. ( Nutanix Portal )

Nutanix documents Storage QoS as the correct mechanism to throttle VM I/O and prevent noisy-neighbor behavior. Nutanix also documents that storage policies in Prism Central apply to Categories, and that QoS settings such as throttled IOPS can be defined in the storage policy itself. That means the administrator can create one policy, set the IOPS cap, and apply it to the category assigned to Dev VMs. Current and future VMs with that category then inherit the same behavior without one-by-one configuration. That is exactly the scalable model described in option B.

The other options do not meet the requirement as cleanly. Replication factor affects resilience, not IOPS governance. Cloning one VM’s QoS is not a robust policy framework for future operations, especially when category-based automation already exists. Compression changes storage efficiency and sometimes performance characteristics, but it does not function as an enforceable per-VM IOPS cap. Because Nutanix explicitly ties QoS throttling to storage policies and category-based assignment, B is the documented and most maintainable answer.

The wording “security drift” points to SCMA rather than AIDE. In Nutanix security hardening guidance, SCMA scheduling is the control that determines how often SCMA runs, with schedule options such as hourly, daily, weekly, and monthly. That makes the SCMA schedule the setting that changes the cadence of security drift checking without changing the underlying controls themselves. ( Nutanix Portal )

AIDE is a file-integrity mechanism and has its own role, but the question is specifically asking about how often the platform checks for drift in a broader security-configuration sense. Changing the list of entities inspected or baseline enforcement behavior would alter scope or enforcement semantics, not merely frequency. The most direct administrative answer is to change the schedule for SCMA. From an exam perspective, Nutanix often separates what runs, how it is enforced, and when it runs into different settings. Here the requirement is only about “how often,” so option B is the precise match. ( Nutanix Portal )

Nutanix documentation for Prism Central point-in-time backup states that Prism Central supports up to 30 days of restore points for recovery. This is documented consistently across Nutanix guidance for backup to Nutanix Objects, AWS S3, and generic S3-compliant storage. Because the platform wording is explicit and repeated, the correct answer is 30 days. ( portal.nutanix.com , portal.nutanix.com , portal.nutanix.com )

This is a retention-boundary question, so it is best answered directly from the official limit rather than from memory. The values 7, 14, and 90 days are exam-style distractors, but Nutanix’s supported restore-point duration for Prism Central recovery is 30 days. In practical terms, that means administrators should plan their backup retention and operational expectations around that documented maximum rather than assuming indefinite or quarterly restore history.

In Nutanix PD-based disaster recovery, when the source storage container name and the destination storage container name are different, the administrator must explicitly configure a vStore mapping. Nutanix documentation says that each vStores entry maps a local source storage container to a remote target storage container. Nutanix also documents that if you want replication from a local vStore to a remote vStore with different names, you provide the mapping between the local and target containers. That is exactly what this question is asking: source container Marketing, destination container Backups, and replication must know how those two names correspond. ( Nutanix Portal )

The other choices do not control storage-container placement for replicated data. Network mapping is about network connectivity for failover or recovery testing, not where the storage data lands. Network compression is unrelated to container targeting. vStore PD is not the Nutanix feature name that solves this requirement. Nutanix’s own remote-container best-practice guidance recommends establishing vStore mappings so the target container is deterministic and properly aligned to the protection design. Therefore, the correct and authentic Nutanix answer is C. vStore mapping. ( Nutanix Portal )

Nutanix volume groups support attaching storage to one or more VMs. Nutanix documentation for creating volume groups states that if you want to attach a volume group to multiple VMs, you select the Shared option. That is the key requirement when building an application-level high-availability configuration where two database VMs need shared access to the same block storage construct. ( Nutanix Portal )

The other options either describe different capabilities or are not the documented mechanism for this requirement. External client access is used when exposing the volume group to external iSCSI initiators, which is not what the question asks. Storing vDisks on different containers does not create shared access semantics. “Multi-writer mode via storage policy” is not the standard Nutanix answer here; Nutanix’s documented configuration language for this use case is the Shared setting on the volume group itself. So if the requirement is two VMs, same application data path, and application-level HA, the correct Nutanix configuration is to create the VG in shared mode. ( Nutanix Portal )

In Prism Central VM efficiency and behavioral-learning views, Nutanix groups inefficient VMs into categories such as over-provisioned, inactive, bully, and constrained. A constrained VM is one that does not have enough resources for its workload and is therefore experiencing pressure. Nutanix documentation ties constrained status to signs such as CPU readiness, memory swapping, or other stress indicators that show the VM is not adequately provisioned for what it is trying to do. That means the best interpretation is that those VMs are lacking sufficient resources, which is option B. ( Nutanix Portal )

This category is different from the other distractors. An inactive VM is mostly idle and not making meaningful use of its assigned resources. An over-provisioned VM is wasting capacity because it was allocated too much. A VM that is merely “consuming too much CPU” could be a bully VM or simply a busy VM, but that alone does not define the constrained classification. Nutanix uses the constrained label to signal a right-sizing opportunity in the direction of adding or adjusting resources, not reclaiming them. Therefore the most accurate answer is B, because constrained VMs are the ones that are short on sufficient resources for their workload behavior. ( Nutanix Portal )

Nutanix documentation confirms that a Prism Element cluster can be registered with only one Prism Central at a time, so option B is false. Nutanix also documents support for deploying multiple Prism Central instances on the same Prism Element cluster, so option A is false as well. The presence of a dedicated Nutanix doc for deploying multiple Prism Central instances on one PE cluster directly disproves option A. ( Nutanix Portal )

That leaves C and D. Nutanix documentation does not position Marketplace as the standard mechanism described in this context for deploying additional Prism Central instances. By contrast, Nutanix clearly documents registration and unregistration workflows, including the fact that clusters can be unregistered when no longer needed and later registered again. Since a cluster can only be attached to one Prism Central at a time, repeating register/unregister operations is a valid administrative action in supported workflows. Therefore D is the best valid statement among the given options. This question is mostly testing whether you know the difference between deployment support and registration exclusivity. Multiple PC instances may exist on the cluster, but registration to management is still one-at-a-time. ( Nutanix Portal )

Here the uploaded answer key says C, but official Nutanix documentation supports D. Nutanix documents that the Recycle Bin retains deleted VMs and volume groups for up to 24 hours by default, and space is not immediately reclaimed while those deleted entities are still being retained. Nutanix also provides a specific workflow to clear space used by the Recycle Bin immediately. That behavior matches the question perfectly: the VMs were deleted earlier today, yet space has not come back. The most likely reason is that the deleted VMs are still in the recycle bin. ( portal.nutanix.com , portal.nutanix.com , portal.nutanix.com )

The answer mentioning “Medusa once every 24 hours” is not the correct interpretation for this symptom. Nutanix’s documented retention mechanism is the recycle bin, not a once-daily reclaim cycle as the main reason space stays consumed immediately after deletion. So based on the official docs, D is the authentic answer.

Here the uploaded answer key says D, but official Nutanix documentation supports B. Nutanix documents that for enhanced security of ROBO environments (typically 1- or 2-node clusters), administrators should select Native KMS (remote) for software-based encryption. Nutanix also states in KMS considerations that software encryption using Native KMS is supported for ROBO deployments using the Native KMS (remote) KMS type. Because the question explicitly describes a 2-node ROBO cluster and requires the keys to be stored outside the cluster, the Nutanix-authentic answer is B. Native KMS (remote).

An external KMS is valid in many enterprise encryption designs, but this question is more specific: it asks for the KMS type Nutanix supports and recommends for software encryption on a ROBO cluster. On that narrow requirement, the official docs point to Native KMS (remote) rather than External KMS. So for an authentic study guide aligned to Nutanix documentation, B is the correct answer.

Here the uploaded answer key says B, but Nutanix documentation recommends at least five stable time sources that have a high degree of accuracy and are reachable over a reliable network. That exact wording appears in Nutanix time synchronization guidance and related KB material. Because the question specifically asks for the minimum number that constitutes a good configuration ensuring a high degree of accuracy, the Nutanix documentation supports 5, which is option C. ( portal.nutanix.com , portal.nutanix.com , portal.nutanix.com )

Three servers is a common generic IT answer, but the Nutanix wording in its own docs is more specific and recommends five. For an NCP-MCI study document meant to stay authentic to Nutanix guidance, C is the correct answer.

Nutanix documentation on capacity reservation explains that a storage container can reserve a portion of storage so that a minimum amount of capacity remains available for that specific container. When reservations are configured, one container may retain protected usable space while other containers in the same storage pool can run out of usable capacity first, even though the pool still shows free space overall. That exactly matches the scenario in the question, which is why C is the correct answer. ( Nutanix )

This is a subtle storage-administration concept. The question is not saying the pool itself is full; it says all but one container are out of space while the single pool still has free space. That pattern is much more consistent with reservation policy than with hardware failure. Failed nodes or drives would usually surface additional health symptoms and would not neatly explain one container still having preserved capacity while the others are exhausted. Nutanix intentionally allows reservation for predictable capacity control, and that is the feature responsible here. Hence A, B, and D are distractors, while C is correct. ( Nutanix )

In Prism Central behavioral learning and VM efficiency views, Nutanix defines an over-provisioned VM as one that is oversized and wasting resources that are not actually needed. If a VM is consistently using less than half of what has been allocated, it is a classic over-provisioning signal: the VM has more CPU and/or memory assigned than its real workload profile requires. ( Nutanix Portal )

The wrong options each represent a different operational pattern. An inactive VM is largely idle. A constrained VM lacks sufficient resources and typically shows signs of pressure, contention, or poor performance. A bully VM consumes disproportionate resources and can negatively affect neighbors. None of those fit the scenario of consistently low utilization relative to allocated sizing. Nutanix includes these VM efficiency categories so administrators can right-size workloads and reclaim capacity before buying more infrastructure. In a resource planning exercise, identifying over-provisioned VMs is especially important because right-sizing them can delay expansion and improve overall efficiency. That is why D is the correct classification. ( Nutanix Portal )

For Prism Central–based DR between on-prem Availability Zones, Nutanix documentation states that the clusters participating in the protection workflow must be registered with Prism Central. In the standard same-management-plane deployment model, the primary and recovery clusters need to be available to the Prism Central instance orchestrating the policy and recovery objects. That is why B is the expected exam answer for successfully configuring asynchronous replication with a protection policy in this scenario. ( Nutanix )

The other options reflect common misconceptions. Identical storage-container naming is not a mandatory prerequisite because Nutanix supports vStore mapping to map source and destination containers when names differ. The strict less than 5 ms RTT requirement is associated with synchronous-style designs such as Metro Availability, not ordinary asynchronous replication. Matching Prism Central versions is good practice but is not the core requirement expressed in the documented prerequisites. Since the protection policy workflow depends on managed registration within the Prism Central DR architecture, B is the best answer. ( Nutanix )

To receive an alert earlier, the administrator must lower the threshold at which the alert’s warning state is triggered. Nutanix alerting works by comparing utilization against configured warning and critical thresholds. If the threshold is lowered, the system reaches it sooner, which gives the administrator more lead time before the environment becomes capacity-constrained. Among the given options, the one that directly performs that function is decreasing the Storage Container Space Usage Exceeded alert Warning threshold, which is option C.

The other choices are weaker or misaligned. There is no meaningful “increase frequency” adjustment that changes when capacity is considered low; it would only change how often checks occur, not the trigger point. Changing the critical threshold does not provide earlier warning notice, which is what the question asks for. The wording also points more naturally to a threshold-based action rather than enabling a different alert family. In Nutanix monitoring, earlier warnings come from making the warning trigger more conservative. Therefore C is the best answer because it causes the alert to fire at a lower usage percentage and provides earlier operational notice.

The requirement is not to keep the mission-critical VMs together or apart from each other; it is to prevent them from running on a specific set of older hosts. In Nutanix Prism Central policy terms, that is a VM-Host Anti-Affinity Policy. Anti-affinity at the VM-host level allows administrators to mark selected hosts as unsuitable targets for selected VMs, which is exactly what the scenario requires. Therefore C is the correct answer.

The distractors all target the wrong relationship. VM-VM affinity and anti-affinity control how VMs are placed relative to other VMs, not whether they can run on particular hosts. VM-Host Affinity would express a positive preference or requirement to run on a chosen host group, but the question is framed negatively: the VMs must never run on the older hosts. Nutanix’s policy model supports this exact exclusion use case through VM-Host Anti-Affinity. In a long refresh delay scenario, this is the safest operational control to keep sensitive workloads off weaker or undesirable hardware without removing those hosts from the cluster immediately. Therefore C is the authentic answer.

The decisive metric in the iostat output is %iowait = 83.20. In Linux performance interpretation, such a high iowait value means the CPU is spending most of its time waiting for disk I/O operations to complete rather than actively executing user or system work. On a Nutanix CVM, that strongly suggests the storage path is the bottleneck at that moment, whether due to backend disk saturation, contention, or broader storage subsystem latency. Since user CPU and steal are both low, the issue is not primarily compute pressure or hypervisor scheduling contention. Therefore, A is the most accurate description. ( Nutanix )

This is a classic Nutanix troubleshooting interpretation question. If the output had shown high %steal, then hypervisor contention would be a stronger candidate. If %system and %user were both very high, compute load from services would be more likely. But with the overwhelming majority of time in iowait, the signal points toward delayed storage completion. In operational terms, the next steps would usually involve checking CVM disk metrics, Stargate performance, storage latency, and possibly the health of the underlying drives. The evidence in the output supports a storage wait condition, which is exactly what option A says. ( Nutanix )