Courseiva

CCNA Vsphere Performance Questions

30 of 105 questions · Page 2/2 · Vsphere Performance topic · Answers revealed

Practice these questionsExam hubAll questions
76
MCQhard

A database VM with 32 vCPUs and 256 GB memory is deployed on a host with two NUMA nodes, each with 8 cores (16 threads with HT). The application performs badly. Analysis shows high remote memory access. What is the MOST LIKELY reason?

A.Hyperthreading is enabled causing CPU contention.
B.The vNUMA nodes are not aligned with the physical NUMA nodes.
C.The host's NUMA nodes are not balanced in memory capacity.
D.The VM has too much memory allocated.
AnswerB

Misalignment causes the guest to access memory from remote NUMA nodes, increasing latency.

Why this answer

With 32 vCPUs, the VM exceeds the default vNUMA node size of 8 cores per NUMA node, so vSphere exposes multiple vNUMA nodes. However, if the vNUMA topology is not aligned with the physical NUMA nodes (e.g., the VM's vCPUs and memory span physical NUMA boundaries), the guest OS may not optimize memory locality, leading to high remote memory access. Option B correctly identifies this vNUMA misalignment as the most likely cause.

Practice this question →
77
Multi-Selecthard

Which THREE techniques are recommended to improve virtual machine network performance in a vSphere environment? (Select three.)

Select 3 answers
A.Enable Jumbo Frames on the virtual switch.
B.Use the VMXNET3 virtual network adapter.
C.Use a dedicated virtual switch for each VM.
D.Use the e1000e network adapter type.
E.Enable SR-IOV on supported NICs.
AnswersA, B, E

Jumbo Frames reduce per-packet CPU overhead, improving throughput.

Why this answer

Correct: A (Enable Jumbo Frames), B (Use VMXNET3), and E (Enable SR-IOV). Jumbo Frames reduce CPU overhead for large packets by allowing larger MTU. VMXNET3 is a paravirtualized adapter that provides high performance and lower latency.

SR-IOV allows a physical NIC to be directly assigned to a VM, bypassing the hypervisor's virtual switch for near-native performance. Option C (dedicated virtual switch per VM) adds complexity without improving performance. Option D (e1000e) is an emulated adapter that is slower than VMXNET3 and should not be used for performance.

Practice this question →
78
Multi-Selecthard

Which THREE factors should be considered when designing a vSphere environment to minimize vCPU overcommitment and improve performance for CPU-intensive workloads?

Select 3 answers
A.Align VM memory with NUMA nodes to reduce remote memory access.
B.Select hosts with a high physical core count per socket.
C.Use VMs with a high number of vCPUs (e.g., 2:1 ratio to pCPUs).
D.Enable Hyper-Threading on all hosts.
E.Use CPU scheduling modes like 'Core Level' to increase performance.
AnswersA, B, E

NUMA-aware placement improves memory access performance.

Why this answer

NUMA-aware VM placement ensures that vCPUs access memory from the local NUMA node, reducing remote memory latency and improving performance for CPU-intensive workloads. Option B is correct because selecting hosts with a high physical core count per socket reduces CPU contention and provides more dedicated resources for each VM. Option E is correct because CPU scheduling modes like 'Core Level' allow vSphere to schedule VMs at the core level, minimizing co-scheduling overhead and improving performance for CPU-bound workloads.

Option C is incorrect because using VMs with a high number of vCPUs (e.g., 2:1 ratio to pCPUs) increases vCPU overcommitment, leading to resource contention and degraded performance for CPU-intensive workloads. Option D is incorrect because enabling Hyper-Threading increases logical CPUs, but for heavily CPU-bound workloads that saturate physical cores, Hyper-Threading can cause resource sharing conflicts and may not improve performance; it can even degrade it due to increased contention for shared execution resources.

Practice this question →
79
MCQhard

A company has a cluster with DRS set to Fully Automated. A VM named VM1 has a reservation of 4 GHz and a limit of 8 GHz on a host with 10 GHz available. The VM is running a batch job that requires as much CPU as possible. The administrator notices that the VM's CPU usage never exceeds 4 GHz. What is the most likely reason?

A.The CPU reservation is preventing the VM from using more than the reserved amount.
B.The CPU shares are too low, causing the VM to be starved.
C.The VM has a CPU limit set to 4 GHz via a resource pool or VM override.
D.DRS is not migrating the VM to a host with more resources.
AnswerC

Correct: A limit caps the maximum CPU usage; the specified limit of 8 GHz may be overridden by a lower limit.

Why this answer

A CPU limit explicitly caps the maximum CPU usage a VM can consume, regardless of available resources. In this scenario, the VM has a reservation of 4 GHz and a limit of 8 GHz on the host, but the administrator observes usage never exceeding 4 GHz, indicating a more restrictive limit is applied via a resource pool or VM override. This overrides the host-level limit, effectively capping the VM at 4 GHz, which matches the observed behavior.

Exam trap

The trap here is that candidates often confuse CPU reservation with CPU limit, assuming a reservation restricts maximum usage, when in fact a reservation only guarantees a minimum and a separate limit must be explicitly set to cap usage.

How to eliminate wrong answers

Option A is wrong because a CPU reservation guarantees a minimum amount of CPU resources to the VM, not a maximum; the VM can burst above the reservation if resources are available, so a reservation of 4 GHz does not prevent usage beyond 4 GHz. Option B is wrong because CPU shares determine relative priority during contention, but since the host has 10 GHz available and the VM is not contending for resources, low shares would not starve the VM; the VM would still be able to use more than 4 GHz if no limit were in place. Option D is wrong because DRS migration is irrelevant when the VM is already on a host with sufficient resources (10 GHz) and the issue is a hard cap on CPU usage, not a lack of available resources on the current host.

Practice this question →
80
MCQeasy

A small financial services company runs a critical trading application on a single vSphere host with two NUMA nodes (each with 8 cores). The VM is configured with 12 vCPUs and 64 GB memory. The application is not yet live, but load testing shows that CPU performance is satisfactory, but memory bandwidth is lower than expected. The application is known to benefit from fast memory access within a single NUMA domain. The administrator notices that the VM's memory is spread across both NUMA nodes. The administrator wants to force the VM to use only one NUMA node to maximize performance. However, the VM requires 12 vCPUs, which exceeds the 8 physical cores per node. What is the best course of action to improve memory bandwidth for this VM?

A.Enable Hyper-Threading and reassign vCPUs to 12, and the VM will automatically run within one NUMA node.
B.Keep the vCPU count at 12 but configure a NUMA Affinity to prefer the first NUMA node, and accept some cross-node memory access.
C.Keep the vCPU count at 12 but set CPU affinity to the first NUMA node only, leaving the second node unused.
D.Reduce the number of vCPUs to 8 to fit within one NUMA node, and keep the same memory size.
AnswerB

NUMA affinity prefers one node for memory allocation, keeping most memory local while allowing vCPUs to use both nodes as needed, balancing performance.

Why this answer

NUMA affinity allows you to prefer a specific NUMA node for memory allocation. Even though the VM has 12 vCPUs exceeding the 8 cores per node, the VM will still run across both nodes for CPU, but memory will be preferentially allocated from the first NUMA node. This reduces cross-node memory access, improving memory bandwidth for the application.

Option A is incorrect because enabling Hyper-Threading does not change the physical core count; 12 vCPUs still exceed 8 cores per node, and the VM cannot be placed entirely within one NUMA node. Option C is incorrect because setting CPU affinity to only the first NUMA node would restrict the VM to 8 cores, causing severe CPU contention and performance degradation. Option D is incorrect; reducing vCPUs to 8 may harm application performance if it requires more CPU resources, and the goal is to improve memory bandwidth while maintaining CPU capacity.

Practice this question →
81
MCQmedium

Based on the exhibit, which statement about the virtual machine's memory state is TRUE?

A.The VM is using the memory balloon driver to reclaim memory.
B.The VM is using transparent page sharing to reduce memory usage.
C.The VM has no memory pressure and is performing optimally.
D.The VM is actively swapping memory because of memory pressure.
AnswerD

SWCUR and swap rates confirm swapping is occurring.

Why this answer

SWCUR (current swapped memory) is 512 MB, and SWR/s and SWW/s show non-zero swap rates, indicating that the VM is actively swapping memory to disk. MCTLSZ is 0, meaning no ballooning is occurring. The VM's memory size is 4096 MB.

Swapping is a sign of memory pressure.

Practice this question →
82
MCQhard

An administrator configures a resource pool with a reservation of 10 GHz and a limit of 20 GHz. The pool contains two VMs: VM1 has a reservation of 6 GHz and no limit, VM2 has no reservation and a limit of 8 GHz. The host has only 12 GHz available. During peak usage, both VMs demand more than their shares. Which statement is true?

A.VM1 gets at least 6 GHz, VM2 gets at most 8 GHz, but because the pool limit is 20 GHz and host only 12, VM2's limit is effectively 6 GHz after VM1's reservation
B.VM1 gets at least 6 GHz, VM2 gets at most 8 GHz, and the remaining 4 GHz is divided by share
C.VM1 gets exactly 6 GHz, VM2 gets exactly 8 GHz, and the pool's reservation is ignored
D.VM1 gets at least 6 GHz, VM2 gets at least 2 GHz due to pool reservation, and both use remaining as share
AnswerA

VM1's reservation takes 6 GHz, leaving 6 GHz for VM2, which is within VM2's limit of 8 GHz, so VM2 gets 6 GHz.

Why this answer

The resource pool has a reservation of 10 GHz and a limit of 20 GHz, but the host only has 12 GHz available. VM1 has a reservation of 6 GHz, so it is guaranteed at least 6 GHz. VM2 has a limit of 8 GHz, which means it cannot exceed 8 GHz.

However, after allocating VM1's reservation, only 6 GHz remains on the host (12 - 6 = 6), which is less than VM2's limit of 8 GHz. Therefore, VM2 effectively gets at most 6 GHz. The pool's reservation of 10 GHz is satisfied because VM1 gets 6 GHz and VM2 gets 6 GHz (6+6=12 >=10).

The pool's limit of 20 GHz is not relevant because the host capacity is the bottleneck. Option B is incorrect because the remaining 0 GHz (12 - 6 - 6 = 0) cannot be divided by share. Option C is incorrect because VM2 cannot get 8 GHz due to insufficient host capacity.

Option D is incorrect because the pool's reservation does not guarantee additional resources beyond individual VM reservations; it only ensures the total allocated to the pool meets that level.

Practice this question →
83
MCQhard

A vSphere cluster uses vSAN with deduplication and compression enabled. An administrator notices that IOPS for a VM is lower than expected despite low CPU and memory utilization on hosts. Which of the following is most likely the cause?

A.The VM storage policy has no read cache reservation
B.vSAN object repair is consuming disk bandwidth
C.The VM is running on a host with a vSAN disk capacity tier that is 80% full
D.Deduplication and compression overhead is causing increased CPU usage on storage controller
AnswerD

Dedupe/compression consumes CPU, and if saturated, can throttle I/O operations.

Why this answer

Deduplication and compression are CPU-intensive operations on vSAN. For workloads requiring high IOPS, the CPU overhead from these features can become a bottleneck, especially if the storage controller or host CPU is saturated, leading to lower IOPS. Option A is incorrect: read cache reservation in a VM storage policy affects read caching but is not a primary cause of low IOPS given low CPU/memory.

Option B is incorrect: vSAN object repair is typically a background task and should not normally impact VM IOPS unless during maintenance. Option C is incorrect: a vSAN disk capacity tier at 80% full does not directly reduce IOPS; vSAN may start rebalancing when near full (above 80% might trigger some activity), but the most likely cause given deduplication and compression enabled is the CPU overhead.

Practice this question →
84
MCQeasy

What is the maximum number of virtual CPUs (vCPUs) that can be allocated to a virtual machine on vSphere 8 Enterprise Plus?

A.128
B.512
C.64
D.256
AnswerD

vSphere 8 supports up to 256 vCPUs per VM.

Why this answer

256 vCPUs. vSphere 8 Enterprise Plus supports up to 256 virtual CPUs per virtual machine. Option A (128) was the maximum in vSphere 6.5 and earlier, not vSphere 8. Option B (512) exceeds the vSphere 8 limit.

Option C (64) is below the current maximum and incorrect.

Practice this question →
85
MCQmedium

An administrator observes that a VM with 4 vCPUs running on a host with hyperthreading enabled shows high %CSTP (co-stop) values in esxtop. What is a likely cause?

A.The VM has CPU affinity configured forcing vCPUs to specific pCPUs.
B.The VM's memory reservation is too high.
C.The host has deep C-states enabled causing CPU idle savings.
D.The host memory is overcommitted causing swapping.
AnswerA

CPU affinity can restrict scheduling and cause co-stop when pCPUs are not all available.

Why this answer

Co-stop occurs when the hypervisor is unable to schedule all vCPUs of a VM simultaneously because the pCPUs are not all available at the same time. This is often due to CPU affinity settings or physical CPU contention. C-state issues and memory pressure do not directly cause co-stop.

Practice this question →
86
MCQeasy

Based on the exhibit, which VM (indicated by the row) is most likely experiencing severe CPU scheduling contention?

A.The VM with %RDY = 2.3
B.The VM with %RDY = 30.0
C.The VM with %RDY = 15.5
D.The VM with %RDY = 0.5
AnswerB

30% ready time indicates severe CPU scheduling contention.

Why this answer

The row with the highest %RDY (30.0) indicates the VM is spending a large percentage of time ready to run but not being scheduled, which is a sign of severe CPU contention. The other rows have lower %RDY values.

Practice this question →
87
MCQhard

A vSphere administrator observes that a VM's memory usage is high, and the host memory is overcommitted. The VM has a memory balloon driver installed and running. Which metric in esxtop would the administrator check to confirm that ballooning is reclaiming memory?

A.SWCUR
B.MCTLSZ
C.ACTV
D.VMMEMCTL
AnswerB

MCTLSZ is the memory balloon size, showing how much memory the balloon driver has reclaimed.

Why this answer

The MCTLSZ field in esxtop's memory section shows the current amount of memory reclaimed by the balloon driver. ACTV shows active memory, SWCUR shows swapped memory, and VMMEMCTL shows memory control overhead. MCTLSZ directly indicates balloon activity.

Practice this question →
88
MCQhard

A vSphere administrator is troubleshooting a performance issue on a VM that has high average latency on its virtual disks. The VM is stored on a VMFS datastore backed by an all-flash array. The VM has 4 vCPUs and 16GB memory. The administrator runs esxtop and sees %USEDRDY: 0.5, %MLMTD: 10, %SWPWT: 0. What is the most likely cause of the high disk latency?

A.Memory ballooning is causing excessive swapping
B.CPU ready time is high
C.Storage controller queue depth is full
D.The VM is experiencing memory contention that is manifesting as disk latency
AnswerD

High %MLMTD (10%) indicates memory contention, which can degrade overall VM performance including disk I/O.

Why this answer

%MLMTD indicates the percentage of time the VM was ready to run but was not given CPU resources due to memory constraints, i.e., memory contention. A value of 10% means the VM is experiencing significant memory throttling, which can cause all I/O operations, including disk, to appear slow. Option D is correct.

Option A is incorrect because %SWPWT is 0, meaning no swapping is occurring, and ballooning would typically lead to swapping. Option B is incorrect because %USEDRDY is only 0.5%, indicating low CPU ready time, not high. Option C is incorrect because there is no metric given (like QFULL) that indicates a full storage queue; the high latency is due to memory pressure, not storage queuing.

Practice this question →
89
MCQmedium

A company runs a four-node vSphere 7 cluster with all-flash storage and 25GbE networking. The cluster hosts 50 VMs, including a critical database VM with 16 vCPUs and 64 GB memory. The database VM's performance has degraded over the past week. vCenter reports high CPU ready time (average 15%) for the database VM, while other VMs show less than 2%. CPU utilization on the host is 70%. The database VM is configured with CPU affinity pinned to cores 0-15 on a dual-socket host with 16 cores per socket (32 total). The administrator also notices that the database VM's CPU co-stop counter is high. What should the administrator do to improve performance?

A.Migrate the database VM to a host with fewer VMs.
B.Enable CPU hot-add for the database VM.
C.Remove the CPU affinity setting for the database VM.
D.Increase the number of vCPUs for the database VM to 32.
AnswerC

Allows the scheduler to balance across NUMA nodes, reducing ready time and co-stop.

Why this answer

The high CPU ready time (15%) and elevated co-stop counter indicate that the database VM is suffering from contention due to CPU affinity pinning to cores 0-15. On a dual-socket host with 16 cores per socket, pinning to cores 0-15 restricts the VM to a single NUMA node, preventing the scheduler from load-balancing across both sockets. Removing the affinity allows the ESXi scheduler to distribute vCPUs across all 32 cores, reducing co-scheduling overhead and ready time.

Exam trap

The trap here is that candidates often assume high ready time means the host is overloaded and choose to migrate or add vCPUs, overlooking that CPU affinity artificially constrains the scheduler and causes co-stop delays.

How to eliminate wrong answers

Option A is wrong because migrating to a host with fewer VMs does not address the root cause—the CPU affinity restriction causing NUMA imbalance and co-stop contention; the current host already has low overall CPU utilization (70%). Option B is wrong because enabling CPU hot-add does not reduce ready time or co-stop counter; it only allows adding vCPUs without reboot, which would likely worsen contention. Option D is wrong because increasing vCPUs to 32 would exacerbate co-scheduling overhead and co-stop delays, especially with affinity pinning to a single socket, leading to further performance degradation.

Practice this question →
90
MCQmedium

A vSphere administrator is troubleshooting a VM that hosts a critical database. The VM is configured with 8 vCPUs and 64 GB memory on an ESXi host with two 10-core CPUs (hyperthreading enabled). Performance charts show CPU ready time averaging 12% during peak hours, and the database application is experiencing high transaction latency. The administrator has verified that no other VMs are contending for CPU. Which action will most likely reduce CPU ready time without negatively impacting database performance?

A.Disable hyperthreading on the host.
B.Assign a CPU affinity mask to pin the VM to specific physical cores.
C.Increase the VM's memory reservation.
D.Reduce the number of vCPUs assigned to the VM to 4.
AnswerD

Reducing vCPUs decreases CPU scheduling overhead and can lower ready time for many applications.

Why this answer

Reducing the number of vCPUs to 4 reduces CPU scheduler overhead and can lower CPU ready time when the VM has more vCPUs than needed. With 8 vCPUs on a host with 20 physical cores (40 logical with HT), and 12% ready time, the VM is likely oversubscribed for its workload. Many database applications do not scale linearly beyond 4 vCPUs, so reducing to 4 often improves performance by decreasing contention and improving cache locality.

Option A is wrong: disabling hyperthreading reduces logical CPUs, potentially increasing contention. Option B is wrong: CPU affinity can lead to unbalanced load and is not recommended in DRS clusters. Option C is wrong: increasing memory reservation does not affect CPU ready time.

Practice this question →
91
MCQeasy

Which tool provides real-time performance monitoring at the ESXi host level, including CPU, memory, network, and storage metrics?

A.dcli
B.esxtop
C.resxtop
D.vCenter performance charts
AnswerB

esxtop runs directly on the ESXi host and shows real-time metrics.

Why this answer

esxtop is a command-line tool that provides real-time host performance data. vCenter performance charts are historical. resxtop is for remote usage, and dcli is a different tool.

Practice this question →
92
MCQmedium

A vSphere environment experiences periodic performance degradation during peak business hours. Analysis shows that one ESXi host's CPU ready time for a specific mission-critical VM is consistently above 20%. Which corrective action should be taken first?

A.Increase the VM's memory reservation
B.Set a higher CPU limit on the VM
C.Migrate the VM to a different host with lower CPU utilization
D.Add more vCPUs to the VM
AnswerC

Moving the VM to a host with more available CPU resources reduces ready time directly.

Why this answer

High ready time indicates CPU contention. Moving the VM to a less loaded host using DRS or manual migration can immediately reduce ready time. Increasing resources without addressing contention is ineffective.

Practice this question →
93
MCQmedium

An administrator configures a resource pool with a CPU reservation and expandable reservation enabled. The root resource pool has 100% of cluster CPU resources. A VM inside the resource pool with no reservation requests CPU resources. What will happen when the VM tries to power on?

A.The VM can power on but will be limited to the resource pool's shares.
B.The VM can power on using resources from the resource pool or its parent due to expandable reservation.
C.The VM cannot power on because the resource pool has a reservation set.
D.The VM can power on using resources from the resource pool only if it has a reservation.
AnswerB

Expandable reservation allows the pool to borrow from parent if needed, so the VM can start.

Why this answer

With expandable reservation enabled, the resource pool can borrow CPU resources from the parent resource pool (root) if needed. Since the VM has no reservation, it does not require any guaranteed CPU capacity; it can power on as long as there are any unreserved resources in the resource pool or parent. Option A is incorrect because shares only affect priority during contention, not the ability to power on.

Option C is incorrect because the reservation on the resource pool does not prevent VMs without reservations from powering on; it only guarantees that a certain amount of resources is reserved for the pool. Option D is incorrect because the VM does not need a reservation to power on; it can use unreserved resources from the pool or borrowed from the parent.

Practice this question →
94
MCQmedium

A vSphere administrator notices that a VM running a critical database is experiencing high latency on its virtual disk. The VM uses a VMDK on a datastore shared by several other VMs. The datastore is backed by an all-flash array. Which action would best reduce latency without increasing storage capacity?

A.Increase the virtual disk shares for the VM
B.Move the VM to another datastore
C.Set a higher I/O limit on the VM
D.Enable Storage I/O Control (SIOC) on the datastore
AnswerD

SIOC uses latency thresholds to throttle I/O and reduce congestion, improving latency for critical VMs.

Why this answer

Storage I/O Control (SIOC) can enforce I/O shares and limits to prioritize latency-sensitive VMs. Increasing shares or limits may not reduce latency if the array is saturated. Moving to another datastore may not be feasible.

Practice this question →
95
MCQmedium

Refer to the exhibit. An administrator runs esxtop and sees the above output for a virtual machine. What is the most likely cause of the performance issue?

A.The VM is running on a host with Hyper-Threading disabled.
B.The memory of the VM is overcommitted.
C.The host has too many other VMs competing for resources.
D.The VM has too many vCPUs for the available physical cores.
E.The VM has CPU affinity configured to a single core.
AnswerD

High %RDY indicates vCPU contention, and high %CSTP indicates co-scheduling overhead, both symptoms of vCPU overcommitment.

Why this answer

The exhibit shows a virtual machine with 8 vCPUs but only 4 physical cores (PCPU) available, as indicated by the %USED and %RUN columns. This leads to CPU ready time (%RDY) being high, which is a classic symptom of CPU overcommitment where the VM's vCPU count exceeds the available physical cores, causing scheduling contention. Option D is correct because the VM has too many vCPUs for the available physical cores, resulting in excessive ready time and degraded performance.

Exam trap

The trap here is that candidates often confuse high %RDY with memory overcommitment or general host contention, but the specific clue is the vCPU count exceeding physical cores, which directly points to CPU overcommitment as the root cause.

How to eliminate wrong answers

Option A is wrong because Hyper-Threading being disabled would reduce the number of logical CPUs but does not directly cause high %RDY; it would instead show lower %CSTP and %MLMTD values. Option B is wrong because memory overcommitment would manifest as high %SWPCT or memory ballooning, not as high CPU ready time in esxtop. Option C is wrong because while other VMs competing for resources can contribute to CPU contention, the exhibit specifically shows a single VM with 8 vCPUs on a host with only 4 physical cores, making the vCPU-to-core ratio the primary issue.

Option E is wrong because CPU affinity to a single core would cause high %CSTP (co-scheduling pressure) and potentially high %MLMTD, not the uniformly high %RDY seen across all vCPUs.

Practice this question →
96
Multi-Selectmedium

Which TWO factors contribute to increased CPU ready time on a vSphere host?

Select 2 answers
A.Memory ballooning due to memory over-commitment.
B.Over-provisioning of vCPUs relative to physical cores.
C.Using CPU affinity to pin VMs to specific cores.
D.Enabling hyper-threading on hosts that already have high vCPU-to-core ratios.
E.High disk latency on the datastore.
AnswersB, D

Too many vCPUs cause contention and ready time.

Why this answer

Over-provisioning vCPUs relative to physical cores leads to contention for CPU resources. When the total number of vCPUs across all powered-on VMs exceeds the number of logical processors (including hyper-threads), the ESXi scheduler must time-share access, resulting in increased ready time as VMs wait for a physical core to become available.

Exam trap

The trap here is that candidates often confuse memory over-commitment (ballooning) with CPU over-commitment, or mistakenly think that CPU affinity always reduces ready time, when in fact it can increase ready time if not used carefully due to load imbalance and reduced scheduler flexibility.

Practice this question →
97
Multi-Selecteasy

A vSphere administrator is scaling a cluster by adding new ESXi hosts and VMs. Which two actions help ensure that performance continues to meet requirements as the environment grows? (Choose two.)

Select 2 answers
A.Place all VMs on the host with the fastest processors.
B.Enable DRS with fully automated mode.
C.Use a vSphere Distributed Switch for network scalability.
D.Configure vSphere HA on the cluster.
E.Set CPU reservations on all VMs.
AnswersB, C

DRS automates initial placement and ongoing load balancing, optimizing performance as VMs are added.

Why this answer

DRS with fully automated mode continuously monitors resource utilization across the cluster and automatically migrates VMs to balance CPU and memory loads. This ensures that as new hosts and VMs are added, workloads are redistributed to prevent hotspots and maintain performance requirements.

Exam trap

The trap here is that candidates often confuse vSphere HA (high availability) with performance scaling, but HA only reacts to failures and does not balance load or optimize resource usage as the environment grows.

Practice this question →
98
Multi-Selecthard

A vSphere cluster is experiencing high CPU ready time across multiple hosts during peak hours. The cluster consists of 8 hosts, each with 2 sockets and 8 cores per socket (hyperthreading enabled). DRS is set to a moderately aggressive migration threshold. The administrator needs to reduce CPU contention without disrupting workloads. Which three actions should the administrator consider? (Choose three.)

Select 3 answers
A.Upgrade the ESXi hosts to models with higher CPU clock speeds.
B.Reduce the number of VMs running on each host.
C.Ensure Hyper-Threading is enabled on all hosts.
D.Configure CPU affinity for VMs with high vCPU counts.
E.Increase the DRS migration threshold to the most aggressive setting.
AnswersA, B, C

Faster CPUs improve per-core performance, reducing the impact of contention.

Why this answer

Enabling hyperthreading increases logical CPUs and can reduce ready time. Reducing the number of VMs per host decreases overall demand. Upgrading to faster CPUs increases per-core capacity.

Setting CPU affinity on VMs is not recommended as it limits DRS balancing. Increasing the DRS migration threshold to the most aggressive may cause excessive migrations and does not address the root cause. Adding vCPUs to VMs would worsen contention.

Practice this question →
99
Multi-Selecteasy

An administrator is analyzing performance data for a vSphere cluster and wants to identify VMs that are experiencing memory pressure. Which two metrics from esxtop or vCenter performance charts reliably indicate that a VM is actively reclaiming memory due to contention? (Choose two.)

Select 2 answers
A.Consumed memory
B.Active memory
C.Ballooned memory
D.Swapped memory
E.Overhead memory
AnswersC, D

Ballooned memory shows the amount of memory reclaimed via the balloon driver, a sign of memory pressure.

Why this answer

Memory ballooning and swapping are clear indicators of memory pressure. Ballooning occurs when the VMkernel reclaims memory from a VM via the balloon driver, and swapping occurs when the hypervisor is forced to swap pages to disk. Active memory is the currently used memory, not a sign of pressure.

Consumed memory includes overhead and target memory. Overhead memory is the hypervisor overhead, not a pressure metric.

Practice this question →
100
Multi-Selecteasy

Which TWO actions are effective in reducing CPU ready time on a vSphere host that is heavily overcommitted on CPU? (Choose two.)

Select 2 answers
A.Add more hosts to the cluster.
B.Enable hyperthreading on the host CPUs.
C.Increase the number of vCPUs on VMs with high ready times.
D.Reduce the number of vCPUs on over-provisioned VMs.
E.Configure CPU affinity to pin vCPUs to specific pCPUs.
AnswersA, D

More hosts spread the load, increasing available pCPUs and reducing contention.

Why this answer

Reducing the number of vCPUs on over-provisioned VMs directly decreases scheduling contention. Adding more hosts to the cluster increases the pool of physical CPUs, reducing overcommitment. Increasing vCPUs worsens the issue.

Enabling hyperthreading does not reduce ready time if the host is already CPU-bound. Configuring CPU affinity restricts scheduling and can increase ready time.

Practice this question →
101
Drag & Dropmedium

Arrange the steps to add an existing virtual machine to a vCenter Server inventory.

Drag steps to the numbered slots on the right, or tap a step then tap a slot.

Steps
Order
1Step 1
2Step 2
3Step 3
4Step 4

Why this order

The correct sequence to add an existing VM to vCenter inventory is: right-click a datacenter or folder, select 'Add Existing Virtual Machine', browse to the VM's .vmx file, and click Finish. This order ensures the wizard is properly initiated and the VM is registered in the desired inventory location.

Practice this question →
102
MCQmedium

A vSphere administrator is troubleshooting a VM that is experiencing excessive disk latency. The VM is on a datastore accessed via NFS over a 1GbE network. The host shows high network utilization. Which action should be taken to improve performance?

A.Enable Storage I/O Control on the datastore.
B.Convert the datastore from NFS to VMFS.
C.Increase the memory reservation for the VM.
D.Upgrade the network link to 10GbE or enable multiple NICs with teaming.
AnswerD

Higher bandwidth reduces latency and congestion.

Why this answer

The VM is experiencing excessive disk latency due to high network utilization on the 1GbE link. Since NFS storage traffic is entirely network-bound, upgrading to 10GbE or enabling multiple NICs with teaming increases the available bandwidth, reduces congestion, and directly addresses the root cause of the latency. This is the most effective action because the bottleneck is at the network layer, not the storage protocol or VM configuration.

Exam trap

The trap here is that candidates may assume Storage I/O Control (SIOC) can solve any storage latency issue, but SIOC only manages contention at the storage array level, not network bandwidth limitations, which is the actual bottleneck in this NFS scenario.

How to eliminate wrong answers

Option A is wrong because Storage I/O Control (SIOC) manages storage queue depth and I/O shares at the datastore level, but it does not increase network bandwidth or resolve a saturated 1GbE link; it is designed for VMFS datastores with multiple VMs contending for storage resources, not for NFS network congestion. Option B is wrong because converting from NFS to VMFS changes the storage protocol but does not increase network throughput; the 1GbE bottleneck remains, and VMFS over iSCSI or Fibre Channel would still suffer from the same network limitation. Option C is wrong because increasing the memory reservation for the VM does not affect network I/O or disk latency; memory reservations guarantee physical RAM but have no impact on storage path bandwidth or network utilization.

Practice this question →
103
MCQhard

Refer to the exhibit. The performance data shows MEMCTL at 5% and SWPOUT at 2%. What does this indicate about the host?

A.The host is over-committed on memory, causing ballooning and swapping.
B.The host is experiencing CPU contention.
C.The host has network congestion.
D.The host has high storage latency.
AnswerA

MEMCTL and SWPOUT indicate memory pressure.

Why this answer

MEMCTL at 5% indicates the ESXi host's memory balloon driver (vmmemctl) is actively reclaiming memory from virtual machines, while SWPOUT at 2% shows the host is swapping guest memory to disk. Together, these values confirm the host is over-committed on memory, forcing the hypervisor to use both ballooning and swapping to free up memory for VMs.

Exam trap

The trap here is that candidates may confuse memory over-commitment indicators (MEMCTL, SWPOUT) with CPU or storage performance metrics, leading them to select CPU contention or storage latency options instead of recognizing the specific memory reclamation counters.

How to eliminate wrong answers

Option B is wrong because CPU contention is measured by metrics like %RDY, %CSTP, or CPU ready time, not MEMCTL or SWPOUT. Option C is wrong because network congestion is indicated by dropped packets, high latency, or errors on virtual switches, not memory-related counters. Option D is wrong because high storage latency is shown by metrics such as KAVG, DAVG, or QAVG in esxtop, not by memory ballooning or swap rates.

Practice this question →
104
Multi-Selecthard

Which THREE vSphere features directly contribute to performance and scaling of virtualized workloads? (Select THREE.)

Select 3 answers
A.vSphere vMotion
B.Storage I/O Control (SIOC)
C.vSphere DRS
D.vSphere Fault Tolerance
E.vSphere HA
AnswersA, B, C

vMotion allows live migration for load balancing and maintenance, aiding scaling.

Why this answer

DRS balances load, SIOC manages storage I/O, and vMotion enables migration for scaling and maintenance. HA provides availability, FT provides fault tolerance, and Backup does not scale performance.

Practice this question →
105
MCQmedium

An administrator observes that a VM with 16 vCPUs on a 2-socket host (8 cores per socket, HT disabled) is experiencing poor performance. The VM's NUMA topology shows 2 vNUMA nodes, but performance is worse than expected. What is the most likely reason?

A.The VM is configured with a single virtual socket with 16 cores, causing the vNUMA nodes to span both physical NUMA nodes and lead to remote memory accesses.
B.The VM's memory is reserved but not backed by physical pages
C.The virtual hardware version is too old
D.The host has insufficient RAM
AnswerA

A single vNUMA node across physical NUMA nodes forces cross-node memory access, increasing latency.

Why this answer

When a VM is configured with a single virtual socket with 16 cores on a dual-socket host (8 cores per physical NUMA node), vSphere creates 2 vNUMA nodes, each with 8 vCPUs. However, the single virtual socket causes vSphere to interleave or misalign the memory of those vNUMA nodes across both physical NUMA nodes, resulting in cross-node memory accesses and degraded performance. Option B is incorrect because memory reservation does guarantee that physical pages are backed; the statement 'memory is reserved but not backed by physical pages' is false.

Option C is incorrect because the virtual hardware version does not affect NUMA topology; the issue is the socket configuration. Option D is incorrect because insufficient RAM would cause memory ballooning or swapping, not the specific cross-node NUMA behavior described.

Practice this question →
← PreviousPage 2 of 2 · 105 questions total

Ready to test yourself?

Try a timed practice session using only Vsphere Performance questions.

Start 20-question session