NCP-AIO Workload Management Practice Question
An MLOps engineer manages a Kubernetes cluster where the NVIDIA GPU Operator runs the MIG manager. Several inference pods must each receive an isolated, fixed slice of a single A100, and the team wants the slices to survive node reboots without manual reconfiguration. Which combination of settings should the engineer apply?
⚠ Common exam trap
The trap here is believing time-slicing delivers the same hardware isolation as MIG, when it only shares one engine among workloads with no memory or fault boundaries.
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 a MIG config labeled for the MIG manager with named profiles and set the device plugin strategy to 'mixed', so whole GPUs and MIG instances are both advertised.
Persistent MIG slices require the MIG manager to read a labeled configuration that defines the desired geometry, which it reapplies after reboots and driver reloads. Setting the device plugin strategy to mixed lets both whole GPUs and the carved MIG instances be advertised under distinct resource names. Pods can then request a specific profile such as nvidia.com/mig-1g.5gb, obtaining fixed, hardware-isolated slices that meet the isolation and persistence requirements.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Set the MIG manager's config to 'all-disabled' and rely on the device plugin to carve profiles dynamically per pod request.
Why it's wrong here
Setting the MIG strategy to all-disabled explicitly prevents MIG instances from being created, so no isolated slices exist for the pods to consume. The device plugin does not carve MIG profiles on demand from a generic request; it only advertises instances that already exist. This configuration therefore delivers whole GPUs rather than the fixed isolated slices the workload requires, and it defeats the isolation goal entirely.
- ✗
Disable the MIG manager entirely and create MIG instances manually with nvidia-smi on each node, then label the nodes so pods schedule there.
Why it's wrong here
Manual nvidia-smi commands do create instances, but they are not reconciled by any controller, so a reboot or driver reload wipes them and pods requesting those resources fail. Node labels alone do not advertise MIG resources to the scheduler; the device plugin must discover the instances. This approach trades automation for fragility and directly contradicts the requirement that slices survive reboots.
- ✗
Enable time-slicing with four replicas and set the MIG manager strategy to 'single', because replicas provide the same isolation as MIG instances.
Why it's wrong here
Time-slicing interleaves workloads on one physical engine with no memory or fault isolation, so it cannot deliver the hard separation the pods require. Combining it with a single MIG strategy also conflicts, since the manager would either apply one geometry or none. The result is shared, best-effort execution rather than the fixed isolated slices the scenario demands, making this combination unsuitable.
- ✓
Configure a MIG config labeled for the MIG manager with named profiles and set the device plugin strategy to 'mixed', so whole GPUs and MIG instances are both advertised.
Why this is correct
A labeled MIG config tells the MIG manager which geometry to apply and persist, so instances are recreated after reboots without manual work. The mixed strategy lets the device plugin advertise both full GPUs and MIG instances, allowing the inference pods to request a specific MIG resource such as nvidia.com/mig-1g.5gb. This yields fixed, isolated slices with hardware-level memory and fault separation.
About these practice questions
This NCP-AIO question is part of Courseiva's 309-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 and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official NVIDIA exam blueprint
This NCP-AIO practice question is part of Courseiva's free NVIDIA 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 NCP-AIO exam.