Free NCP-MCI 7.5 practice test: 30 questions with answers

Foundation images the nodes with AOS and the chosen hypervisor and then creates the cluster. Life Cycle Manager updates a cluster that already exists, Move migrates workloads, and Prism Central manages clusters only after they are registered.

The replication factor, RF2 or RF3, sets how many redundant copies of each block of data are stored, which determines how many simultaneous node or disk failures the cluster can survive. VM placement limits, compression, and NFS client access are controlled by other settings.

A scale-out Prism Central runs multiple management VMs so the service continues if one of them fails. Snapshots of a single VM support recovery but not continuous availability, and a workstation or Prism Element-to-Prism Element registration does not add Prism Central redundancy.

LACP requires the switch to form a matching link aggregation group on those ports, whereas an active-backup bond needs no such switch configuration. Per-VM VLANs and routing protocols are not requirements for an LACP host bond.

AHV connects VM NICs to VLAN networks through Open vSwitch, and active-backup is a valid uplink bond mode in which one link is active. LACP is not the only failover option, VM networks do not attach to the CVM management IP, and VMs do not need dedicated physical NICs.

Erasure coding protects data with parity fragments, which uses less capacity than extra full copies and suits cold data. Replication factor 3 increases overhead, while snapshot schedules and Metro Availability address recovery and site protection rather than capacity efficiency.

Cluster Lockdown restricts cluster CLI access to the SSH public keys an administrator has added. Flow controls VM network traffic, data-at-rest encryption protects stored data, and Move handles workload migration.

The cluster virtual IP gives administrators a single fixed address for Prism, and the cluster serves it from whichever node is currently leading Prism. CVM addresses identify individual controller VMs, gateways route VM traffic, and live migration uses host-level networking.

A VM-host affinity policy restricts a VM to the hosts listed, which is what this scenario needs. Anti-affinity separates VMs from each other, a category is only a label used for grouping, and snapshots restore data rather than control host placement.

A container whitelist defines the client IP addresses or subnets allowed to access the container over NFS. Replication factor, compression, and erasure coding govern data protection and efficiency, not client access.

Nutanix Volumes presents block storage through volume groups over iSCSI, which suits guest clustering. Files provides SMB or NFS file shares, Objects provides S3-compatible buckets, and protection domains replicate data rather than presenting shared block devices.

AHV presents virtual disks and NICs through VirtIO devices, so Windows guests need the VirtIO drivers to use them. The other options describe requirements that do not apply when a Windows VM is migrated to AHV.

Flow security policies can select VMs by category, so new or moved VMs carrying the same category are covered automatically. Static IP lists must be maintained by hand, protection domains manage replication, and volume groups present storage.

Nutanix Guest Tools enables application-consistent snapshots for supported Windows applications through VSS. VirtIO includes storage drivers as well as network drivers, Prism Central is managed from outside the guest, and Foundation is for imaging nodes.

VM high availability restarts affected VMs on other hosts, and anti-affinity spreads an application's VMs across hosts so one host failure does not take all of them down. Disabling live migration, enabling compression, and pinning every VM to one host do not improve host-failure tolerance.

Cloning creates a new VM that is independent of the source, so changes to the clone do not affect the original. Live migration moves a VM without creating a copy, snapshot restores revert the original VM, and affinity policies control placement only.

Protection domains group VMs and volume groups so they share snapshot schedules and replicate to a remote site together. Recovery plans belong to Prism Central-based DR, categories are only tags, and a storage container holds data rather than defining replication groups.

Recovery plans in Prism Central orchestrate failover and failback between availability zones, including ordering and testing. Protection domains are a Prism Element construct, a witness only arbitrates Metro-style failover, and the Controller VM does not control replication schedules.

Synchronous replication writes to both sites before acknowledging, which gives zero data loss when the link can support it. Scheduled asynchronous replication leaves a recovery point gap, and snapshots on the source cluster or configuration exports do not protect against loss of that site.

Prism Central holds management configuration and data for the registered clusters, so a backup kept off the instance is what allows it to be restored. Encryption, cluster counts, and alert settings do not make Prism Central recoverable.

A remote site entry on the source Prism Element cluster identifies the destination and how to reach it, which replication requires. Recovery plans are a Prism Central construct, categories are tags, and volume groups store data rather than define destinations.

The witness provides an independent tie-breaker so that only one site continues as the active copy, which prevents split-brain. The witness does not hold the only copy of the data, replace the replication link, or supply management addresses.

Local snapshots help with rollback but live on the same cluster, so losing that cluster removes them, while replicated recovery points can be used at a remote site. Snapshots do not replace replication, replication does not require power-off, and snapshots can restore VM disks.

Capacity planning in Prism Central projects how long current resources will last based on usage trends. Health checks report problems, Life Cycle Manager handles updates, and Pulse sends telemetry to Nutanix support rather than forecasting capacity.

N+1 sizing keeps enough free CPU and memory on the remaining hosts so that the VMs from a failed host can restart on them. Replication factor one removes data redundancy, disabling HA removes the restart behaviour the goal depends on, and powering off a CVM takes away storage services.

Removing a node moves its data and VMs onto the remaining nodes, so the cluster must have enough capacity and redundancy to stay protected afterward. Clock settings, container ownership, and the cluster virtual IP are not removal prerequisites.

Analysis charts in Prism show performance metrics such as CPU, memory, and I/O over time for entities, so heavy consumers can be spotted. Foundation, Move, and the image upload wizard do not provide performance trend analysis.

Life Cycle Manager inventories the components and performs updates in a controlled, rolling sequence on a running cluster. Foundation is for initial imaging, NCC runs health checks rather than updates, and Objects is storage, not an update engine.

Maintenance mode evacuates the VMs from the host, normally through live migration, so the reboot causes no VM downtime. Deleting the virtual switch, shutting VMs down manually, or disabling HA does not prepare the host safely.

Prism alerts and NCC health checks both surface cluster health problems and recommended actions. Move, Foundation, and Objects bucket policies serve migration, imaging, and object access purposes rather than general health reporting.