I85PREP
⌘ K
HCIP-Cloud Computing study materials
HCIP-Cloud Computing study resources

HCIP-Cloud Computing practice questions

One question at a time. Build your HCIP-Cloud Computing understanding.

First 50 of 200 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1Computing

In Huawei FusionCompute, which component provides the centralized web-based management portal used to create clusters, hosts, and virtual machines?

  1. FusionCompute VRM Correct answer.
  2. FusionCompute CNA
  3. UltraVR
  4. FusionStorage Manager

Explanation

The Virtual Resource Manager (VRM) is the management node that exposes the FusionCompute web portal and holds the cluster, host, and VM configuration data. The CNA is the compute agent node that actually runs virtual machines. FusionStorage Manager and UltraVR serve storage management and disaster recovery roles respectively.

Question 2Computing

Which Huawei virtualization technology allows a running virtual machine to be migrated between hosts with no noticeable service interruption?

  1. Clone
  2. Snapshot
  3. Live migration Correct answer.
  4. HA restart

Explanation

Live migration moves a running VM's memory and state to another host while the guest continues executing, so downtime is negligible. Snapshot and clone create copies of disk state but do not relocate a running VM. HA restart only restarts a VM after a host failure, which causes an outage.

Question 3Computing

A FusionCompute cluster contains hosts with different CPU models from the same vendor. VMs must be able to live migrate freely among all hosts. What should the administrator configure?

  1. Larger memory reservations per VM
  2. Per-VM CPU affinity to a single host
  3. A separate port group per host
  4. A baseline CPU model for the cluster Correct answer.

Explanation

Setting a cluster baseline CPU model masks newer instruction sets so that a VM started on any host can run on hosts with older CPUs, enabling unrestricted live migration. Memory reservations, port groups, and CPU affinity do not address instruction-set compatibility between hosts.

Question 4Computing

Which FusionCompute feature groups multiple physical NICs on a host so that traffic continues if one link fails?

  1. VLAN trunking
  2. Bonding of physical NICs Correct answer.
  3. Memory ballooning
  4. Thin provisioning

Explanation

NIC bonding aggregates physical interfaces into a logical bond, providing redundancy and, depending on mode, load sharing so traffic survives a link failure. VLAN trunking only tags traffic, memory ballooning reclaims RAM, and thin provisioning allocates storage on demand.

Question 5Computing

An administrator needs a VM's virtual disk to consume storage only as data is actually written, rather than reserving the full size at creation. Which disk type meets this requirement?

  1. Thick-provisioned disk
  2. Preallocated disk
  3. Thin-provisioned disk Correct answer.
  4. Raw device mapping

Explanation

A thin-provisioned disk allocates physical capacity on demand as blocks are written, so initial consumption is small. Thick or preallocated disks reserve the full declared size immediately. Raw device mapping exposes a LUN directly and is not a thin allocation model.

Question 6Computing

Which FusionCompute mechanism lets a VM keep its network identity when it is migrated to a host in a different subnet?

  1. CPU QoS capping
  2. Disk thin provisioning
  3. Snapshot rollback
  4. Migration across Layer 2 or Layer 3 networks with remapping Correct answer.

Explanation

FusionCompute supports cross-subnet migration by handling the address translation or network remapping needed so the VM's connectivity and identity persist after moving to a host on a different Layer 3 segment. Snapshots, thin provisioning, and CPU QoS have no bearing on network identity during migration.

Question 7Computing

In FusionCompute, what is the primary purpose of configuring a distributed virtual switch rather than relying only on host-level virtual switches?

  1. To deduplicate storage blocks across hosts
  2. To increase the physical CPU frequency of hosts
  3. To encrypt VM memory at rest
  4. To apply consistent network policy across many hosts from one control point Correct answer.

Explanation

A distributed virtual switch centralizes port groups and security or traffic policies so that consistent networking applies to VMs regardless of which host they run on. It does not change CPU frequency, deduplicate storage, or encrypt memory.

Question 8Computing

A VM template must be deployed many times with each copy retaining the same OS configuration but independent disks. Which operation should be used?

  1. Take a snapshot of the source VM
  2. Deploy from template with full copy Correct answer.
  3. Attach the source VM disk in shared mode
  4. Convert the source VM to a raw device mapping

Explanation

Deploying from a template with a full copy produces independent VMs with their own disks based on the template configuration. Snapshots preserve state for rollback rather than producing independent copies, shared disks are for clustering, and raw device mapping exposes a LUN directly.

Question 9Computing

Which statement correctly describes the role of the FusionCompute CNA?

  1. It runs on each physical host and provides the virtualization layer that hosts VMs Correct answer.
  2. It brokers authentication for external LDAP users
  3. It provides block storage deduplication for the storage pool
  4. It stores cluster metadata and serves the management web portal

Explanation

The Compute Node Agent (CNA) is installed on each physical server and supplies the hypervisor layer that executes virtual machines. Cluster metadata and the web portal belong to the VRM, deduplication belongs to the storage layer, and directory authentication is handled by the management plane.

Question 10Computing

An administrator wants automatic restart of VMs on surviving hosts when a physical host fails. Which capability must be enabled?

  1. Live migration scheduling
  2. Memory overcommit
  3. Cluster high availability Correct answer.
  4. Disk thin provisioning

Explanation

Cluster HA monitors hosts and restarts affected VMs on remaining hosts after a failure. Memory overcommit affects consolidation ratios, thin provisioning affects storage allocation, and live migration scheduling only moves running VMs under normal operation.

Question 11Computing

What does a memory reservation on a virtual machine guarantee in FusionCompute?

  1. The storage IOPS ceiling for the VM's disk
  2. The number of virtual NICs assigned to the VM
  3. A minimum amount of physical memory that the host keeps available for that VM Correct answer.
  4. A maximum CPU frequency the VM may reach

Explanation

A memory reservation sets a floor of physical memory the host must keep for the VM, protecting it from ballooning and host-level contention. CPU frequency limits, NIC counts, and storage IOPS ceilings are configured by separate CPU, network, and storage QoS settings.

Question 12Computing

A host is under memory pressure and the administrator observes that the hypervisor is reclaiming guest memory pages that the guest believes it still owns. Which technique is being used?

  1. NIC bonding
  2. Memory ballooning Correct answer.
  3. Disk snapshot
  4. CPU affinity

Explanation

Ballooning uses a driver inside the guest to inflate a balloon that reserves guest pages, letting the hypervisor reclaim physical memory without swapping. CPU affinity, snapshots, and NIC bonding operate on processor placement, disk state, and network links respectively.

Question 13Computing

Which FusionCompute setting limits the maximum number of IOPS a virtual disk may issue?

  1. Memory reservation
  2. Disk QoS upper limit Correct answer.
  3. VLAN priority
  4. CPU shares

Explanation

Disk QoS settings let an administrator cap or guarantee IOPS and throughput for a virtual disk. Memory reservation, CPU shares, and VLAN priority govern memory, processor scheduling, and network tagging, not storage IOPS.

Question 14Computing

An administrator must protect a VM so it can be restored to an earlier point after a failed software upgrade. Which action is appropriate?

  1. Increase the memory reservation
  2. Bond additional physical NICs
  3. Enable cluster HA
  4. Create a snapshot before the upgrade Correct answer.

Explanation

A snapshot records the VM disk state at a point in time so the VM can be rolled back if the upgrade fails. Cluster HA, memory reservation, and NIC bonding address host failure recovery, memory guarantees, and link redundancy, none of which restore prior disk state.

Question 15Computing

Which statement best explains why a VM's virtual disk should be placed on a datastore with sufficient free capacity before creating a memory snapshot?

  1. Snapshot delta files and memory state are written to the datastore, so insufficient space can cause the snapshot to fail Correct answer.
  2. Snapshots are stored only in host RAM and never touch storage
  3. Snapshot creation requires the datastore to be thick-provisioned
  4. Snapshots are written to the VM's network port group

Explanation

When a snapshot is taken, changed disk blocks and, if included, memory state are written as delta files on the datastore. If free space is inadequate the snapshot operation can fail or the VM can be affected. Snapshots are not held solely in RAM, do not require thick provisioning, and are never stored on a port group.

Question 16Computing

What is the effect of enabling memory overcommit on a FusionCompute cluster?

  1. Virtual disks are automatically deduplicated
  2. Each VM is guaranteed dedicated physical cores
  3. VM network traffic is encrypted in transit
  4. The sum of configured VM memory may exceed physical host memory Correct answer.

Explanation

Memory overcommit allows the total configured memory of powered-on VMs to exceed the physical memory of the host, relying on reclamation techniques such as ballooning and swapping. It does not guarantee dedicated cores, deduplicate disks, or encrypt network traffic.

Question 17Computing

An administrator wants to move a VM's disk files to a different datastore without powering off the VM. Which operation should be used?

  1. Snapshot deletion
  2. HA restart
  3. Storage live migration Correct answer.
  4. Cold migration

Explanation

Storage live migration relocates virtual disk files between datastores while the VM remains running. Cold migration requires the VM to be powered off, snapshot deletion removes delta files, and HA restart recovers VMs after failure rather than relocating storage.

Question 18Computing

Which FusionCompute practice best reduces the risk of a single physical switch failure isolating a host's management traffic?

  1. Place management traffic on the same bond as storage traffic only
  2. Assign all management traffic to one NIC on one switch
  3. Connect management NICs to separate physical switches in a redundant pair Correct answer.
  4. Disable the management VLAN

Explanation

Spreading management interfaces across two independent physical switches removes a single point of failure in the network path. Concentrating management traffic on one NIC or one switch preserves the risk, disabling the VLAN breaks separation, and mixing management with storage on one bond does not add switch-level redundancy.

Question 19Computing

What does the FusionCompute concept of a resource cluster primarily provide?

  1. A physical rack of storage arrays
  2. A set of VLAN tags for tenant isolation
  3. A license key that enables live migration
  4. A logical grouping of hosts whose resources are pooled for scheduling and HA Correct answer.

Explanation

A cluster groups hosts so their CPU and memory are pooled, enabling scheduling, load balancing, and HA across the group. It is not a storage rack, a VLAN tag set, or a license artifact.

Question 20Computing

An administrator observes that a VM cannot start because the cluster lacks sufficient memory reservation capacity, even though free physical memory appears adequate. Which explanation is most likely?

  1. The VM's virtual NIC is disconnected
  2. The VM has too many snapshots
  3. The datastore is thin-provisioned
  4. Admission control checks reserved capacity, and existing reservations consume the reservable pool Correct answer.

Explanation

Admission control evaluates whether the requested reservation can be satisfied from the reservable capacity of the cluster; other VMs' reservations can exhaust that pool even when raw free memory looks sufficient. NIC state, thin provisioning, and snapshot count do not govern reservation admission.

Question 21Computing

Which FusionCompute feature supports running a VM with a virtual disk stored on a distributed storage pool rather than a local disk?

  1. Memory ballooning
  2. FusionStorage integration Correct answer.
  3. VLAN trunking
  4. CPU affinity

Explanation

FusionStorage provides a distributed storage pool that FusionCompute hosts can use as a shared datastore, enabling features such as live migration without shared external arrays. CPU affinity, ballooning, and VLAN trunking are unrelated to storage placement.

Question 22Computing

A tenant requires that VMs belonging to different departments cannot see each other's broadcast traffic on the same physical network. Which configuration achieves this?

  1. Increase disk QoS for each department
  2. Disable HA on the cluster
  3. Enable memory overcommit for all VMs
  4. Place each department in a separate VLAN-backed port group Correct answer.

Explanation

Separate VLAN-backed port groups isolate broadcast domains so departments cannot see each other's Layer 2 traffic. Memory overcommit, disk QoS, and HA settings affect resource allocation and recovery, not network isolation.

Question 23Computing

Which statement accurately describes the relationship between a VM snapshot and the base disk?

  1. The snapshot replaces the base disk permanently
  2. Writes after the snapshot go to a delta file while the base disk is preserved for rollback Correct answer.
  3. The base disk is deleted immediately when a snapshot is created
  4. Snapshots convert the base disk from thin to thick provisioning

Explanation

After a snapshot, new writes are redirected to a delta file while the original base disk remains unchanged, allowing rollback to the captured state. The base disk is not deleted or replaced, and snapshots do not change the provisioning type of the base disk.

Question 24Computing

Which FusionCompute capability allows a VM to be started on a different host automatically after its original host becomes unreachable?

  1. VLAN trunking
  2. Cluster HA with host isolation response Correct answer.
  3. Memory reservation
  4. Thin provisioning

Explanation

Cluster HA detects host failure or isolation and restarts the affected VMs on remaining hosts according to the isolation response policy. Thin provisioning, VLAN trunking, and memory reservation do not provide failure-triggered restart.

Question 25Computing

An administrator plans to reduce storage consumption for many similar VMs. Which combination is most appropriate?

  1. Linked clones from a base image plus thin-provisioned disks Correct answer.
  2. Raw device mappings plus snapshots
  3. Memory overcommit plus NIC bonding
  4. Thick-provisioned disks plus full clones

Explanation

Linked clones share the base image's disk blocks and thin provisioning allocates only written blocks, together minimizing storage use for similar VMs. Thick disks with full clones maximize consumption, raw mappings bypass the datastore model, and memory overcommit and NIC bonding do not affect storage footprint.

Question 26Computing

In FusionCompute, which component runs on each physical host and provides the virtualization layer that executes virtual machines?

  1. The VNC proxy
  2. The VRM
  3. The FusionStorage client
  4. The CNA agent Correct answer.

Explanation

The Compute Node Agent (CNA) is installed on each physical host and supplies the virtualization platform on which guest VMs run, while also reporting to the management component. The VRM is the centralized management node and does not execute guest workloads itself.

Question 27Computing

Which FusionCompute component provides the centralized web-based management plane for clusters, hosts, and VMs?

  1. The hypervisor kernel
  2. The VRM Correct answer.
  3. The CNA
  4. The storage multipath driver

Explanation

The Virtualization Resource Manager (VRM) is the management node that exposes the management interface, stores configuration, and coordinates cluster operations. Hosts are managed by it but do not themselves provide the central management plane.

Question 28Computing

A FusionCompute cluster must keep VMs running if one physical host fails. Which cluster capability provides this?

  1. Template deployment
  2. Memory overcommit
  3. High availability (HA) Correct answer.
  4. Dynamic resource scheduling (DRS)

Explanation

Cluster HA monitors hosts and restarts affected VMs on surviving hosts when a failure is detected, which is exactly the continuity requirement described. DRS balances load across healthy hosts but does not by itself recover VMs from a failed host.

Question 29Computing

Which statement describes the purpose of a virtual switch in FusionCompute?

  1. It provides block storage to hosts over iSCSI
  2. It connects virtual NICs to physical uplinks and enforces port-level network policy Correct answer.
  3. It replicates VM disks to a remote site
  4. It schedules vCPU time across physical cores

Explanation

A virtual switch is the software switching layer that attaches VM virtual NICs to physical NICs and applies policies such as VLAN tagging and security settings. It is a network construct and does not perform storage, CPU scheduling, or replication duties.

Question 30Computing

What is the role of a VM template in FusionCompute?

  1. It is a policy object that defines HA restart priority
  2. It is a running VM that other VMs use as a shared disk
  3. It is a backup copy used only for restoring a failed host
  4. It is a master image from which new VMs with identical base configuration are deployed Correct answer.

Explanation

A template captures a preconfigured OS and software baseline so that many VMs can be deployed consistently and quickly. Deployed VMs receive their own virtual disks and do not share a running master as a disk.

Question 31Computing

Which FusionCompute object is the logical container that groups hosts so that HA, DRS, and shared policies can apply across them?

  1. The cluster Correct answer.
  2. The datastore
  3. The virtual NIC
  4. The port group

Explanation

A cluster groups physical hosts under shared scheduling and availability policies, which is why HA and DRS are configured at cluster scope. Port groups, datastores, and virtual NICs are narrower objects that do not aggregate hosts.

Question 32Computing

In FusionCompute, which virtual disk type allocates storage space only as guest data is actually written?

  1. Thick-provisioned disk
  2. Thin-provisioned disk Correct answer.
  3. Shared read-only disk
  4. Raw device mapping

Explanation

Thin provisioning allocates physical capacity on demand, so the disk consumes only what the guest writes rather than its full configured size. Thick provisioning reserves the full size at creation time.

Question 33Computing

Which FusionCompute feature lets a VM's virtual disk reside on a shared distributed storage pool accessible to multiple hosts?

  1. Virtual disk on a shared datastore Correct answer.
  2. Virtual disk on a USB passthrough device
  3. Virtual disk on a local host disk
  4. Virtual disk on a CD-ROM image

Explanation

Placing a virtual disk on a shared datastore makes it reachable from every host in the cluster, which is a prerequisite for migration and HA restart. Local disks, CD-ROM images, and USB passthrough do not provide cluster-wide disk access.

Question 34Computing

An administrator needs a VM's virtual disk to be reachable from every host in a cluster so the VM can be restarted elsewhere after a host failure. Which storage placement satisfies this?

  1. A virtual floppy image stored on the management node
  2. A datastore created from one host's local RAID volume
  3. A datastore backed by shared storage presented to all hosts Correct answer.
  4. A temporary disk on the host's ramdisk

Explanation

HA restart requires the VM's disk to be accessible from the surviving host, so the datastore must be shared across cluster hosts. Host-local storage would leave the disk unreachable after failure.

Question 35Computing

A VM's guest OS reports low free space while the datastore still shows ample capacity. Which explanation is most consistent with thin-provisioned disks?

  1. The datastore has silently converted the disk to thick provisioning
  2. The guest filesystem has filled its configured virtual size even though physical blocks are allocated lazily Correct answer.
  3. The hypervisor has compressed the disk, hiding free space
  4. The VM is using raw device mapping instead of a virtual disk

Explanation

With thin provisioning, the guest sees the full configured virtual size, and running out of space inside the guest is independent of remaining physical capacity on the datastore. The two capacities are tracked separately, so a full guest filesystem can coexist with free datastore space.

Question 36Computing

An administrator must prevent a production VM from being migrated automatically by DRS during business hours. Which action addresses this?

  1. Remove the VM from its port group
  2. Set the VM's automation level to manual or disabled for DRS Correct answer.
  3. Change the VM's virtual disk to thick provisioning
  4. Disable HA on the cluster

Explanation

DRS migration decisions honor per-VM automation settings, so setting the VM to manual or disabled keeps it from being moved automatically while still allowing explicit migrations. Cluster-wide HA and storage settings do not control DRS placement.

Question 37Computing

A VM must communicate on VLAN 20 while its host's physical uplink carries multiple VLANs. What must be configured on the virtual switch port group?

  1. Bind the VM to a dedicated physical NIC
  2. Set the port group's VLAN ID to 20 Correct answer.
  3. Set the VM's virtual NIC MTU to 20
  4. Enable promiscuous mode on the port group

Explanation

The port group's VLAN ID determines which tagged VLAN the virtual switch associates with frames from VMs attached to it. Tagging is handled at the port group level, not by MTU or by dedicating a physical NIC.

Question 38Computing

After a host failure, HA attempts to restart a VM but the operation fails. Which condition most plausibly explains this?

  1. The VM's guest OS is a 64-bit build
  2. The VM's disk resides on a datastore not accessible to the surviving hosts Correct answer.
  3. The cluster has DRS set to fully automated
  4. The VM has more vCPUs than the failed host had cores

Explanation

HA restart requires the surviving host to open the VM's disk, so a datastore that is not shared with that host blocks the restart. vCPU counts, guest bitness, and DRS automation do not prevent a restart on their own.

Question 39Computing

Which FusionCompute practice improves availability of the management plane itself?

  1. Place all hosts' management NICs on a single physical switch
  2. Deploy the VRM in a redundant configuration across separate hosts Correct answer.
  3. Run the VRM on the same host as all production VMs
  4. Store VRM backups only on the VRM's own system disk

Explanation

Redundant VRM nodes on separate hosts let management survive the loss of one node or host. Co-locating everything or keeping backups only on the failing node removes that protection.

Question 40Computing

An administrator wants to reduce storage used by many near-identical VMs after deployment. Which approach is most appropriate?

  1. Disable thin provisioning on the datastore
  2. Give each VM a raw device mapping LUN
  3. Deploy from a common template and enable linked clones or deduplication where supported Correct answer.
  4. Convert every VM disk to thick provisioning

Explanation

Deploying from a shared template with linked clones or deduplication lets many VMs share unchanged base blocks instead of storing full copies. Thick provisioning and dedicated LUNs increase consumption rather than reduce it.

Question 41Computing

A VM's virtual disk is on a datastore that is running low on free space. Which risk is most immediate?

  1. The VM's MAC address will change
  2. The VM's vCPU count will be automatically reduced
  3. The VM may pause or fail to write because the thin disk cannot grow Correct answer.
  4. The VM will be migrated to another cluster

Explanation

Thin disks grow as data is written, so exhausting the datastore can halt writes and may pause the VM. CPU count and MAC addressing are unaffected by datastore capacity.

Question 42Computing

Which statement correctly distinguishes a snapshot from a backup in FusionCompute?

  1. A snapshot captures a point-in-time VM state on the same storage, while a backup is an independent copy that survives loss of that storage Correct answer.
  2. A snapshot replaces the need for a backup because it is immutable
  3. A backup is only a record of VM configuration without disk data
  4. A snapshot is stored offsite, while a backup stays on the same datastore

Explanation

Snapshots preserve a rollback point but typically depend on the original storage, so losing the datastore can lose the snapshot too. A backup is a separate copy that remains usable if the primary storage fails.

Question 43Computing

An administrator must move a powered-on VM to another host with no interruption to its network sessions. Which condition is required?

  1. The source and destination hosts share access to the VM's datastore and compatible CPU features Correct answer.
  2. The VM has an attached USB passthrough device
  3. The VM uses a local disk on the source host
  4. The destination host has a different virtual switch name

Explanation

Live migration needs the destination host to reach the same storage and to present a compatible CPU feature set so the running VM can continue. Local-only disks, mismatched networking, and host-bound passthrough devices prevent an uninterrupted move.

Question 44Computing

Which FusionCompute setting should be reviewed when a VM's clock drifts badly under heavy load and time-sensitive applications misbehave?

  1. The cluster's DRS threshold
  2. The datastore's block size
  3. The VM's time synchronization configuration with the host Correct answer.
  4. The VM's disk provisioning type

Explanation

Guest time keeping depends on synchronization with a reliable time source such as the host, so reviewing time synchronization addresses drift. Disk provisioning, DRS thresholds, and block size do not govern guest clock accuracy.

Question 45Computing

A cluster uses memory overcommit and a host becomes memory-stressed. Which outcome should the administrator expect?

  1. Each VM's configured memory is permanently reduced
  2. VMs are automatically converted to templates
  3. The host's physical memory is expanded dynamically
  4. The hypervisor reclaims memory through ballooning or swapping, which can degrade guest performance Correct answer.

Explanation

Overcommit relies on reclaim techniques such as ballooning and swapping when physical memory is short, and those techniques cost performance. Configured memory values are not permanently changed and physical memory is not expanded at runtime.

Question 46Computing

An administrator must ensure a latency-sensitive database VM always has a guaranteed share of physical memory. Which setting addresses this?

  1. Lower the VM's CPU limit
  2. Enable memory overcommit on the cluster
  3. Configure a memory reservation for the VM Correct answer.
  4. Set the VM's disk to thin provisioning

Explanation

A memory reservation guarantees that the specified amount of physical memory remains available to the VM and is not reclaimed. Overcommit and CPU limits do not provide that guarantee.

Question 47Computing

A VM needs to reach a storage LUN directly for an application that requires raw device access. Which configuration matches this requirement?

  1. A linked clone of the VM
  2. A thin-provisioned virtual disk on a shared datastore
  3. A snapshot of the VM's system disk
  4. Raw device mapping of the LUN to the VM Correct answer.

Explanation

Raw device mapping presents a physical LUN directly to the guest so the application can use raw access semantics. Virtual disks, snapshots, and linked clones are abstracted storage objects and do not provide direct LUN access.

Question 48Computing

A cluster's hosts have different CPU generations. Live migration of a VM between them fails. Which action most directly resolves this?

  1. Configure a cluster CPU baseline that masks features not common to all hosts Correct answer.
  2. Convert the VM's disks to thick provisioning
  3. Increase the VM's memory reservation
  4. Disable HA on the cluster

Explanation

A cluster CPU baseline exposes only features present on all hosts, giving a compatible CPU model for migration. Memory, disk provisioning, and HA settings do not affect CPU feature compatibility.

Question 49Computing

An administrator observes that a VM cannot be restarted by HA because the cluster lacks sufficient reserved memory on surviving hosts, even though total configured memory looks adequate. Which factor best explains this?

  1. HA admission control checks reservable capacity, and reservations or overcommit reduce what is actually available Correct answer.
  2. HA ignores memory and only checks free disk space
  3. The cluster's DRS threshold is set too low
  4. The VM's MAC address conflicts with another VM

Explanation

Admission control evaluates whether surviving hosts can actually guarantee the resources needed, so existing reservations and overcommit reduce the reservable pool below the nominal total. Disk space, MAC addressing, and DRS thresholds do not determine memory admission.

Question 50Computing

A tenant requires that VMs in two departments cannot see each other's broadcast traffic while sharing the same physical uplinks. Which design satisfies this?

  1. Place each department in a separate VLAN-backed port group on the shared virtual switch Correct answer.
  2. Disable the virtual switch and use only physical NICs
  3. Assign both departments the same VLAN ID but different IP subnets
  4. Attach both departments to the same port group and rely on guest firewalls

Explanation

Separate VLAN-backed port groups isolate broadcast domains at layer 2 even when the same physical uplinks are used. Sharing a port group, sharing a VLAN, or removing the virtual switch does not provide that isolation.