Q: 6
How can you run applications on GPU worker nodes in clusters created using Container Engine for
Kubernetes (OKE)?
Options
Discussion
Option C, D is tricky, but it’s about specifying GPU, not CPU, in the pod spec. I’ve seen similar practice questions that try to mess you up with this wording. Pretty sure C is what Oracle wants.
D talks about CPU resource requests, but if the question asked for 'CPU' not 'GPU' in the spec would your pick change?
Why does B mention including CUDA in the app container rather than on the node image?
C makes sense, but B is tempting if you think about using virtual node pools with custom containers. Pretty sure they expect C though since it covers managed node pool, GPU shape, and CUDA pre-installed. Anyone else leaning toward B for anything?
Is there any case where D would be right when the question clearly asks about GPU in the pod spec?
Yeah, it's definitely C for this case.
Don't think D fits here, the trap is specifying CPU not GPU in pod spec. C is right.
Be respectful. No spam.
