VCP-DCV vSphere Performance and Scaling Practice Question
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?
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Keep the vCPU count at 12 but configure a NUMA Affinity to prefer the first NUMA node, and accept some cross-node memory access.
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.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Enable Hyper-Threading and reassign vCPUs to 12, and the VM will automatically run within one NUMA node.
Why it's wrong here
Hyper-Threading does not change the 8-core limit per NUMA node; the VM still cannot fit entirely in one node.
- ✓
Keep the vCPU count at 12 but configure a NUMA Affinity to prefer the first NUMA node, and accept some cross-node memory access.
Why this is correct
NUMA affinity prefers one node for memory allocation, keeping most memory local while allowing vCPUs to use both nodes as needed, balancing performance.
- ✗
Keep the vCPU count at 12 but set CPU affinity to the first NUMA node only, leaving the second node unused.
Why it's wrong here
CPU affinity to only 8 cores would cause severe CPU contention and high ready time for the 12 vCPUs.
- ✗
Reduce the number of vCPUs to 8 to fit within one NUMA node, and keep the same memory size.
Why it's wrong here
Reducing vCPUs may negatively impact application performance if the application requires 12 vCPUs.
Go deeper
Related to this question
About these practice questions
This VCP-DCV question is part of Courseiva's 498-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This VCP-DCV practice question is part of Courseiva's free VMware certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the VCP-DCV exam.