VCP-DCV vSphere Performance and Scaling Practice Question
A company has a vSphere 7 cluster with 6 hosts, each with 2 sockets (16 cores per socket, hyperthreading enabled) and 1 TB RAM. The cluster uses DRS with a migration threshold of 4 (moderate). A single VM runs a latency-sensitive trading platform with 12 vCPUs and 256 GB RAM. During market hours, the administrator notices that the VM's CPU ready time spikes to 15%, and the application reports high response times. The administrator runs esxtop on the host and sees the following: %RDY for the VM is 15%, %CSTP is low, and the host's overall CPU utilization is 60%. The VM is on a host that also has 5 other VMs with smaller resource footprints. The administrator wants to minimize disruption while improving performance. Which action should the administrator take first?
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
✓
Configure the VM's NUMA node preference to use a single NUMA node.
Trading VMs with high vCPU counts often benefit from NUMA-awareness. The host has 2 NUMA nodes (one per socket). With 12 vCPUs, the VM likely spans both nodes, causing remote memory access. By setting the VM to prefer one NUMA node (or limiting vCPUs to 8), performance improves. Option D is the most effective immediate step. Option A may cause excessive vMotion. Option B might not address the root cause. Option C would reduce parallelism and might not help if the application is multithreaded.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Manually migrate the VM to a host with fewer running VMs.
Why it's wrong here
The host has only 5 other VMs; CPU utilization is only 60%, so contention is likely not due to too many VMs.
- ✗
Increase the VM's CPU reservations to guarantee CPU cycles.
Why it's wrong here
Reservations ensure minimum CPU but do not reduce ready time if the VM is waiting on memory stalls.
- ✗
Reduce the VM's vCPU count from 12 to 8.
Why it's wrong here
This could help, but the application may require 12 vCPUs; a more targeted NUMA adjustment is better first.
- ✓
Configure the VM's NUMA node preference to use a single NUMA node.
Why this is correct
This reduces remote memory access, which is a common cause of high ready time in large VMs.
Go deeper
Related to this question
About these practice questions
Courseiva writes every VCP-DCV question from scratch — 498 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 →
Same concept, more angles
1 more way this is tested on VCP-DCV
These questions test the same concept from different angles. Work through them to make sure you can recognise it however the exam phrases it.
Variation 1. A financial services company has a vSphere cluster with four hosts, each equipped with 2 CPU sockets (10 cores each, hyperthreading enabled) and 512 GB RAM. The cluster runs a mix of production and test VMs. Recently, a critical trading application VM (8 vCPUs, 64 GB RAM) was moved by DRS to a host that also runs an analytics workload (4 vCPUs, 32 GB RAM). After the move, the trading application's performance degraded significantly. The administrator checks vCenter performance charts and sees that the trading VM's CPU ready time has increased from 2% to 20% since the migration, while the host's overall CPU utilization is only 50%. The analytics VM shows normal performance. What is the most likely cause of the performance degradation?
easy- A.The host has NUMA node interleaving enabled in BIOS.
- ✓ B.The trading VM is spanning multiple NUMA nodes on the destination host.
- C.The trading VM's CPU affinity is incorrectly set after migration.
- D.The host has insufficient memory to run both VMs, causing memory overcommitment.
Why B: The trading VM has 8 vCPUs, and the destination host has two NUMA nodes (one per socket). With hyperthreading, each node has 10 cores × 2 threads = 20 logical CPUs. An 8-vCPU VM may be scheduled across both NUMA nodes, causing remote memory access latency when a vCPU accesses memory that is local to the other node. This latency can manifest as increased CPU ready time (stalls) and degraded performance. The analytics VM with 4 vCPUs likely fits within one node, so its performance remains normal. Option B correctly identifies NUMA spanning. Option A is less plausible because NUMA interleaving is a BIOS setting that would affect all VMs, not just after migration. Option C is unlikely because CPU affinity is not automatically changed by DRS. Option D is incorrect because memory overcommitment is not present (host has 512GB, VMs use only 96GB).
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.