Question 22
UnclassifiedWhat is the purpose of the NVIDIA GPU device plugin for Kubernetes?
Correct answer: B
Explanation
The NVIDIA GPU device plugin registers GPU capacity with Kubernetes so the scheduler can see and allocate it to pods. That is why pods can request the extended resource "nvidia.com/gpu" in their resource limits and receive access to NVIDIA GPUs.
Why each option is right or wrong
A. It installs NVIDIA drivers on Kubernetes nodes automatically
B. It exposes GPU resources to Kubernetes so pods can request 'nvidia.com/gpu' in resource limits
Under the Kubernetes device plugin framework (Kubernetes API machinery for extended resources), a device plugin advertises allocatable hardware to the kubelet and registers it as an extended resource that the scheduler can place on a pod. In NVIDIA’s implementation, that resource is exposed under the name `nvidia.com/gpu`, so a pod can request it in `resources.limits` and be assigned GPU devices by the kubelet at admission time.
C. It provides a web dashboard for monitoring GPU utilization in the cluster
D. It manages GPU firmware updates across Kubernetes nodes