hardMultiple Choice
350-401 Practice Question: Is troubleshooting a VMware vSphere cluster where…
A network engineer is troubleshooting a VMware vSphere cluster where a VM with a large memory footprint (256 GB) is experiencing poor performance. The host has two NUMA nodes, each with 128 GB of memory. The VM is configured with 256 GB of memory and 4 vCPUs. Performance monitoring shows high memory latency and CPU ready time. What is the most likely cause?
⚠ Common exam trap
Cisco often tests the misconception that memory performance issues are always due to overcommitment or ballooning, but the trap here is that the VM's memory size exactly matches the total host memory, leading candidates to overlook the NUMA boundary constraint.
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
✓
The VM's memory size forces it to span multiple NUMA nodes, increasing memory access latency.
The VM is configured with 256 GB of memory, but each NUMA node on the host has only 128 GB. Since a single NUMA node cannot satisfy the VM's memory allocation, the hypervisor must split the VM across both NUMA nodes. This forces memory accesses to cross the NUMA interconnect (e.g., QPI or UPI), which introduces significantly higher latency compared to local memory access, directly causing the observed high memory latency and increased CPU ready time.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
The VM's memory size forces it to span multiple NUMA nodes, increasing memory access latency.
Why this is correct
On a NUMA (Non-Uniform Memory Access) system, each node contains local memory with low latency, while accessing memory from a remote node traverses the QPI/UPI interconnect, adding significant latency. If a VM's memory allocation exceeds the capacity of a single NUMA node, the hypervisor must back the guest's physical memory with pages from multiple nodes, forcing some accesses to be remote. This VM-to-NUMA-node mismatch directly increases memory access latency and is the correct explanation, because the symptom is memory latency, not CPU or host pressure.
- ✗
The VM has too few vCPUs for the memory size.
Why it's wrong here
The number of vCPUs assigned to a VM determines its CPU scheduling capacity and can affect application throughput or context-switch overhead, but it has no impact on the physical path or speed of memory access. Even if the VM has too few vCPUs to efficiently drive its workload, that would manifest as CPU starvation or run-queue delays, not as higher memory access latency in a shared NUMA topology. Thus, this option confuses compute resource allocation with the memory subsystem's NUMA locality, making it an incorrect cause for the described latency problem.
- ✗
The host is using memory ballooning to reclaim memory from other VMs.
Why it's wrong here
Memory ballooning is a host-side reclamation technique where a balloon driver inside the guest inflates to push unused pages back to the hypervisor, typically in response to memory overcommitment. While ballooning can force a guest to reclaim or swap pages if it becomes memory-constrained, it does not change the physical NUMA node placement of the VM's memory or the latency of accesses to already-assigned local memory. Since the problem statement is specifically about memory access latency caused by exceeding a NUMA node, ballooning is a separate issue related to host memory pressure, not the VM's memory topology.
- ✗
The host's memory is overcommitted.
Why it's wrong here
The host has 256 GB total memory, and the VM's allocation matches that value, meaning there is no memory overcommitment for this particular VM; the hypervisor can fully back the guest's memory without reclaiming pages. Overcommitment would only arise if the sum of all VM memory reservations exceeded physical RAM, and even then it would lead to swapping or ballooning, not to NUMA span or remote memory access. Moreover, overcommitment affects availability and performance under pressure, but it does not directly force the VM's memory to be placed across multiple NUMA nodes, so it is an incorrect explanation for the latency issue.
Go deeper
Related to this question
About these practice questions
Courseiva writes every 350-401 question from scratch — 1,923 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or 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 350-401 practice question is part of Courseiva's free Cisco 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 350-401 exam.