An AKS node pool needs to be dedicated to GPU workloads only. How do you prevent CPU workloads from being scheduled there?

1. A Set node pool as system type
2. ✓ B Apply a taint (gpu=true:NoSchedule) and add matching tolerations to GPU pod specs (correct answer)
3. C Use a NetworkPolicy
4. D Set maxPods to 1

Shortcut: Tainting the GPU node pool with NoSchedule prevents non-GPU pods (which lack the toleration) from being scheduled there. GPU pods must have the matching toleration.