NCP-MCI-7.5 Nutanix Certified Professional - Multicloud Infrastructure (NCP-MCI) 7.5 Questions and Answers

Nutanix documentation for Active Directory authentication in Prism Element states that the default LDAP port is 389, while port 636 is used for LDAPS, which provides encrypted LDAP traffic. Because the question explicitly says the administrator wants all traffic encrypted, the correct port is 636. That is the standard LDAPS answer and the one Nutanix documents directly.

The missing exhibit is not actually necessary to answer the question because the encryption requirement itself is enough. Ports 443 and 9440 are associated with web interfaces or management access, not LDAP directory authentication transport. Port 389 is unencrypted LDAP unless other mechanisms are layered on top. Nutanix’s own directory-authentication docs make the intended answer clear: use LDAPS on port 636. Therefore C is the authentic Nutanix answer.

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 )

The requirement is very specific: the organization wants a single place to review security posture and compliance across on-prem Nutanix plus public cloud environments, and the output should support security operations workflows, not just basic cluster health views. Nutanix documents Security Central as the product designed for exactly this purpose. The official guide states that Security Central provides a unified view of security posture across Nutanix private cloud and AWS and Azure public clouds, and it is positioned as a multi-cloud security and compliance capability rather than just a local Prism view.

The other answers are narrower than what the question asks. Flow Network Security policy views focus on segmentation and policy state, not full multi-cloud posture management. Prism Central alerts and events are useful operational signals, but they do not replace a posture-and-compliance platform. The standard Prism Central security dashboard is not the same as Security Central’s broader cloud-security scope. Because Nutanix explicitly documents Security Central as covering Nutanix, AWS, and Azure with posture and compliance visibility, D is the authentic answer.

Nutanix networking guidance makes this distinction very clearly: balance-slb (Active-Active with MAC pinning) is the bond type used when the administrator wants multi-uplink utilization without additional physical-switch configuration. Nutanix also documents that LACP requires Active-Active / balance-tcp, and balance-slb must not be combined with link aggregation protocols such as LACP. If the switch team turns on LACP while the Nutanix side remains in balance-slb, the uplink behavior is mismatched and degraded throughput is the expected result. Therefore D is the correct answer.

This is a classic exam trap because both balance-slb and balance-tcp are Active-Active styles, but they are not interchangeable. One is switch-independent and the other is switch-dependent with LACP. The problem here is not MTU, uplink count, or support for vs0. It is the fact that the switch configuration and AHV bond mode no longer agree. In Nutanix, if LACP is enabled on the switches, the virtual switch must be configured for the LACP-compatible Active-Active mode, not MAC pinning. That is why D is the authentic answer.

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 )

The requirement is straightforward: there is one source site (Site A) and the workload must be replicated to two destination clusters, each with potentially different RPO settings. In Nutanix DR design terminology, that is a one-to-many replication pattern. Nutanix documentation for multisite disaster recovery explains that a single protection policy can include replication schedules to multiple recovery clusters, which is how a source can protect workloads to more than one target with different schedule definitions. That makes D the correct answer.

The other choices describe different topologies. Many-to-One means multiple sources targeting one destination. Many-to-Many is broader bidirectional or multiple-source/multiple-target logic than the question requires. Two-Way-Mirroring is not the correct pattern for one source sending copies outward to two other sites. Nutanix’s protection-policy model is specifically flexible enough to define different replication schedules toward separate destinations, which is why this exact scenario is best labeled One-to-Many. So for an exam-aligned and documentation-consistent answer, D is the right choice.

Nutanix documentation for Erasure Coding is explicit: a cluster must have a minimum of four nodes to enable erasure coding, and for RF2 the minimum remains four nodes. Nutanix further explains that RF3 with erasure coding requires six nodes, which helps distinguish the two thresholds clearly for exam purposes. Since the question asks specifically about Replication Factor 2 and Erasure Coding, the minimum supported node count is 4. ( Nutanix )

This question is testing one of the most common node-count rules in Nutanix storage efficiency. A three-node cluster can provide RF2 storage, but that does not automatically mean erasure coding is available there. Erasure coding introduces its own resiliency and stripe-width requirements, and Nutanix sets the minimum at four nodes for RF2. The values 5 and 6 are distractors meant to confuse RF2 and RF3 rules. Because the official requirement is four nodes for RF2 EC, the correct answer is B. ( Nutanix )

Nutanix documentation specifically identifies deduplication as providing the greatest benefit for common data sets, and it explicitly calls out full-clone VDI workloads as one of the strongest matches. In a full-clone desktop environment, many virtual desktops contain highly similar operating-system files, application binaries, and repeated data blocks. Capacity deduplication is designed to eliminate those repeated blocks and therefore reduce the total physical storage consumed by the environment. This is exactly why Nutanix recommends deduplication for such use cases.

The other answers are less targeted. Thin provisioning is useful generally but is not the signature optimization for repeated-content VDI storage. Erasure coding is powerful for write-cold data and capacity efficiency, but it is not the primary answer Nutanix recommends for full-clone VDI patterns. “Post-process map reduce” is not the standard storage-efficiency answer expected in this context. For exam purposes, whenever Nutanix asks about full-clone VDI, the important association is with capacity deduplication, because that feature directly exploits the high block similarity across cloned desktops. Therefore B is the correct and most authentic answer.

Nutanix documentation for Updating Capacity Configurations explains that the Reserve Capacity For Failure option determines whether runway analysis should account for the impact of a node failure, and when Auto Detect is selected, Prism Central automatically uses the cluster’s failure assumptions rather than the full unconstrained cluster total. In practical runway terms, that means the calculation reflects the loss of the node whose failure would have the greatest effect on usable capacity. Among the provided options, the best match for that documented logic is D: the runway analysis excludes the resources of the largest node so estimates account for a node failure. ( Nutanix Portal )

This behavior is important because runway is intended to answer “how long until I run out of safe usable capacity,” not just “how long until all raw resources are consumed.” Nutanix therefore lets administrators model capacity while preserving headroom for failure scenarios. The fixed 10% answer is too generic and not how Nutanix documents the option. Storage-only reservation also misses the fact that runway tracks CPU, memory, and storage dimensions. So, based on Nutanix’s Auto Detect wording and the resilience-oriented design of runway planning, D is the most authentic answer. ( Nutanix Portal )

Nutanix documentation clearly states that guest VMs with GPU passthrough cannot be migrated to other hosts in the cluster. By contrast, Nutanix does support live migration for some vGPU- enabled VMs under supported conditions, but not for passthrough GPUs. That distinction is exactly what this question is testing. Therefore the correct answer is D. ( portal.nutanix.com , portal.nutanix.com , portal.nutanix.com )

This is a high-value exam distinction: vGPU and GPU passthrough are not interchangeable in migration behavior. Memory overcommitment and vNUMA may affect placement or performance, but they are not the documented hard stop presented here. If a VM uses GPU passthrough, live migration is not supported, so the administrator cannot evacuate it the way they can with ordinary VMs.

Nutanix Intelligent Operations documentation for behavioral learning and VM right-sizing identifies several thresholds used to detect resource pressure, including CPU ready time > 5% and memory swap rate > 0 Kbps. Those are direct indicators that the VM is experiencing contention rather than simply being oversized or idle. A VM in that state is not comfortably provisioned; it is operating under resource pressure. Nutanix uses the term constrained VM for workloads that are lacking sufficient resources and therefore show signs of stress such as CPU wait, memory swapping, or storage-path pressure. That makes C the correct answer.

The high Host I/O Stargate CPU usage strengthens the case that this is an under-served or stressed workload rather than an inactive or over-provisioned one. An inactive VM would not show these pressure signals. An over-provisioned VM typically wastes allocated resources but does not struggle for them. A bully VM consumes disproportionate resources and affects others, but the threshold clues given in Nutanix right-sizing guidance map more directly to a VM that is constrained. Therefore the most accurate category for these observed characteristics is C. Constrained VM.

Nutanix security and network-segmentation documentation clearly states that the segmented DSIP and the regular cluster DSIP are different entities, and that Prism Central needs a cluster DSIP in the same subnet to upgrade. In other words, placing the DSIP only in a secondary non-routable segment can break Prism Central’s ability to perform upgrades if Prism Central cannot reach the cluster using the expected subnet path. The documented fix is to maintain or configure a DSIP in the original communication subnet that Prism Central can use. That makes C the correct answer.

The other options are not the Nutanix-documented solution. Proxy ARP is not the prescribed remedy here, and disabling segmentation is unnecessary if the communication requirement can be satisfied by adding the proper DSIP. Uplink redundancy has nothing to do with the management-plane reachability issue in the question. So the authentic fix is to configure an additional DSIP in the original subnet used for Prism Central communication, which is C.

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.

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.

Nutanix Move documentation states that when the administrator changes the preparation mode for only a subset of VMs under Change Settings, and there is now a mix of preparation modes across the migration plan, the overall preparation mode automatically switches to Mixed. That is exactly the condition described in the question: 10 VMs in the migration plan, two manually changed, which creates more than one preparation mode across the plan. Therefore the correct answer is D. Mixed. ( Nutanix Portal )

This is an important Nutanix Move behavior because it affects how the migration plan is displayed and managed. The administrator does not get a brand-new custom label such as “Hybrid”; Nutanix specifically uses the term Mixed to indicate that some VMs are prepared one way and others another way. It also means the plan must be reviewed carefully so each VM’s preparation setting aligns with guest OS requirements and migration prerequisites. Since the official Move documentation names this exact automatic state change, the authentic answer is D rather than Manual or Automatic. ( Nutanix Portal )

Nutanix documentation for enabling Marketplace in Prism Central explicitly states that a unique data services IP address (DSIP) must be configured on the Prism Element cluster that hosts Prism Central. This is a hard prerequisite for Marketplace enablement and application deployment workflows. In exam terms, the presence of DSIP is not optional convenience; it is a setup requirement Nutanix calls out directly before Marketplace can be used. ( Nutanix )

The other answers are not Marketplace prerequisites in the same direct sense. SMTP is useful for email notifications, LDAP for external authentication, and LACP for network design, but none of those is the documented enabling requirement for Marketplace. Nutanix separates platform application management prerequisites from general infrastructure best practices, and DSIP falls into the former category. So when the exam asks what must already be in place on the Prism Element cluster before enabling Marketplace, the correct answer is A: Data services IP must be configured. ( Nutanix )

The key clue is the mismatch between the VM being slow under memory overcommit and the guest OS pagefile usage remaining low. If the guest itself were under strong memory pressure, you would expect higher pagefile or swap-file activity inside the OS. Instead, the symptoms suggest memory reclamation is happening outside the guest, which points to hypervisor swapping. In Nutanix AHV terms, that means the platform is reclaiming memory at the virtualization layer, causing performance degradation even though the guest does not show heavy internal paging. Therefore C is the most likely cause.

The distractors are weaker. “CVM swapping” would indicate pressure in the storage/control path VM, not necessarily this guest’s memory experience. Ballooning and swapping are not mutually exclusive, and the question’s evidence does not specifically prove ballooning is disabled. This is a troubleshooting pattern-recognition question: slow VM, full memory use, low guest pagefile use, and overcommit enabled usually indicates the hypervisor is swapping, so C is the best answer.

Incrementing rx_crc_errors usually point to a Layer 1 or physical-path problem, not a CPU or memory issue. CRC errors mean frames are arriving corrupted, and the most common reasons are a faulty cable, damaged transceiver, bad port, or another physical medium issue between the AHV host and the switch. Because the CRC counter is specifically increasing while other counters like missed, FIFO, and general rx errors remain low, this strongly suggests corruption during transmission rather than host overload. That makes B the correct answer.

A duplex mismatch can cause broader link-performance problems, but the pattern of steadily rising CRC errors is much more classically tied to bad physical media. MTU issues usually show fragmentation or packet-handling problems, not CRC corruption itself. CPU saturation does not create on-the-wire checksum corruption. In Nutanix troubleshooting, when CRC counters rise repeatedly, the correct operational response is to inspect and swap the cable or optics, verify switch-port health, and test the NIC path physically. So B is the strongest technical match.

Nutanix guidance for compression settings and current container behavior aligns with inline compression as the standard default, with an immediate (0-minute) delay model. Nutanix compression FAQ guidance also states that when deploying compression on a container, the recommended default delay for inline compression is 0, reinforcing the expected default posture for new storage containers. That makes A the correct answer. ( Nutanix )

This is one of those questions where older platform knowledge can cause confusion because historical versions and terminology evolved. The safest exam-oriented approach is to follow the current documented behavior and Nutanix’s own recommended default delay. Post-process compression with 60 minutes used to be a common mental model in older storage discussions, but the modern expected answer for a newly created container in Prism Central is inline compression with a 0-minute delay. Therefore A is the right choice. ( Nutanix )

Nutanix separates storage resiliency reservation from compute high availability. For storage, the relevant setting is Rebuild Capacity Reservation, which reserves storage space so the cluster can self-heal and rebuild after node, block, or rack failure. For memory and compute, the relevant concept is High Availability, which reserves or accounts for host resources needed to restart workloads after a host failure. Putting those together gives the combination that best governs the remaining usable storage and memory capacity while preserving fault tolerance. That makes C the most correct answer.

The distractors either use non-standard phrasing or split the concepts incorrectly. “Memory Reservation” by itself is not the standard Nutanix answer when the exam is asking about cluster-wide restart capacity and usable memory after failure accounting. Nutanix administrators think in terms of HA for compute-memory protection and rebuild capacity reservation for storage protection. If the goal is to manage the usable remaining capacity for both storage and memory correctly, the right pair is Rebuild Capacity Reservation and High Availability. Therefore C is the authentic answer.

Nutanix documentation is very clear on this point: Secure Boot is not supported if the VM uses IDE disks. Nutanix also states that for AHV, if Secure Boot is enabled for UEFI mode, the IDE interface is not supported and SATA should be used for the VirtIO ISO in that configuration. So an IDE-attached disk is a direct blocker to enabling Secure Boot. ( Nutanix Portal )

This is one of those areas where the exam expects you to know compatibility rules, not just feature names. UEFI is actually part of the supported Secure Boot path, so option D is not the blocker. The CVM’s RAM and node cache storage are unrelated to whether the VM’s firmware and bus configuration can support Secure Boot. The administrative fix is usually to convert the VM to a supported configuration using UEFI and non-IDE bus types where required. Therefore, among the listed options, the condition that prevents completion of the task is the presence of IDE disks, making C correct. ( 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.

Nutanix Volumes documentation explains that external iSCSI clients use the iSCSI Data Services IP address for target discovery and connection. Nutanix specifically documents creating or updating the ISCSI Data Services IP Address on the cluster as part of enabling access to Volumes. If that step is missing, an external initiator can fail to discover or connect to the volume group correctly. Therefore B is the correct answer.

The other options are unrelated to the direct connectivity requirement. Microservices Infrastructure and Prism Central registration are not the necessary cluster-side prerequisites for an external iSCSI connection. LDAP directory configuration is also not the primary dependency being tested here. The key Nutanix platform requirement is that the cluster exposes the Volumes service via the configured Data Services IP. Without that, the initiator connection workflow breaks. So B is the authentic answer.

In Prism Central, Runway Analysis is not a simple reading of current utilization; it is a forecast based on historical consumption trends. Nutanix documentation explains that capacity planning and runway use machine learning and historical resource behavior to estimate when CPU, memory, or storage will run out. That means a cluster can still show only 60% memory used today and yet have a short memory runway if the memory-consumption trend has been climbing steadily and rapidly over recent history. In this scenario, the most likely reason for only 12 days of runway is sustained growth in memory use over the historical observation period. ( Nutanix )

The distractors do not fit the documented behavior as well. Nutanix does support scenario modeling, but a low runway in the standard view is primarily a forecast outcome, not a default penalty for missing history or a result of guaranteed HA mode by itself. HA reservations can reduce available effective capacity, but the wording of the question points to forecasting logic rather than static reservation math. Since Nutanix runway is based on actual consumption trends, the answer that best explains a low runway despite moderate present-day utilization is that memory usage has been continuously increasing, which is option B. ( Nutanix )

Nutanix documentation for creating a storage container with Replication Factor 1 shows the workflow in the Prism Element web console. The official steps begin with logging on to Prism Element, navigating to storage, and creating the RF1-enabled container there. This makes A the correct answer.

This question is really about administrative scope. RF1 is a specialized storage-container setting that is managed at the cluster level, and Nutanix documents it in Prism Element rather than Marketplace or Self-Service. Prism Central can manage many cluster-wide resources, but the specific documented RF1 container-creation workflow is rooted in Prism Element. For exam prep, when a question asks which interface is required to create an RF1 container, the safe and documentation-aligned answer is Prism Element. Therefore A is the authentic Nutanix 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 )

Here the uploaded answer key says C, but current Nutanix documentation says the opposite. Nutanix’s Live vDisk Migration Across Storage Containers documentation explicitly states that during an ongoing vDisk migration you cannot perform operations such as clone the VM, resize the VM, or take a snapshot. Nutanix also states, under DR considerations, that you should remove the VM from the protection domain or protection policy, migrate the vDisks, and then add the VM back afterward. That means the actions listed in option C are specifically blocked during vDisk live migration. ( Nutanix Portal )

Because all four answer choices are problematic against the current documentation, this is one of the questions I would flag as flawed in the source bank. The most accurate real-world Nutanix answer is: you should not perform snapshot, clone, resize, or DR policy/protection-domain operations during the migration. So although the uploaded key marks C, the official docs indicate that C is not actually allowed. For your final study document, this question should ideally be corrected or rewritten because the documented AHV behavior does not cleanly support any listed option as a valid “can perform during migration” task. ( Nutanix Portal )

Nutanix documentation for PD-based DR remote-site configuration states that if vStore mapping is not configured, replication occurs to the remote container with the same name as the source container. Nutanix further explains that this fallback relies on the existence of that matching destination container. Therefore, if same-name containers do not exist on the two sites and no explicit vStore mapping is defined, the replication target cannot be resolved correctly and the replication will fail. That makes A the correct answer.

This behavior is intentional. Nutanix does not randomly pick a destination container, nor does it silently force all replications into the default container. When administrators want different container names between source and target, the supported mechanism is vStore mapping. Skipping that mapping only works when identical container names already exist at both sites. Because the question explicitly removes that condition, replications fail to complete, which is why A is the authentic answer.

Nutanix documentation for Protection Policies supports protecting entities by category, which is the most scalable and administratively efficient method when many current and future VMs share the same RPO requirement. By assigning the marketing VMs to a category and then associating that category with the policy, the administrator avoids one-by-one maintenance and also ensures new VMs can inherit protection simply by category assignment. ( Nutanix Portal )

This is exactly the kind of pattern Nutanix encourages across Prism Central: use categories for dynamic grouping and policy application. Adding each VM individually works, but it is not the easiest or most maintainable approach for a large VM set. Templates are not the unit used here for attaching data protection to running workloads. A category-based protection model gives consistency, scale, and easier future operations, particularly when teams like Marketing may continuously create or receive new VMs. Since the question asks for the easiest way, the category-based workflow in option A is the cleanest answer and aligns directly with official Nutanix policy design. ( Nutanix Portal )

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.