Scheduling

CKA v1.35 · Workloads and Scheduling (15%) → "Configure Pod admission and scheduling (limits, node affinity, etc.)" Target time: 12 minutes

Node muscle-worker2 has been labelled disktype=ssd and tainted workload=batch:NoSchedule.

In namespace sched, create a Deployment batch (3 replicas, image nginx:1.27-alpine) that:

1. Tolerates the workload=batch:NoSchedule taint.
2. REQUIRES a node with disktype=ssd (hard requirement, not a preference).
3. Spreads its pods across nodes with topologySpreadConstraints: maxSkew 1, topologyKey kubernetes.io/hostname, whenUnsatisfiable ScheduleAnyway.
4. Declares requests cpu=50m memory=32Mi.

All 3 replicas must be Running, and all 3 must land on muscle-worker2.

NOTE: this scenario taints muscle-worker2. Left in place, that taint stops other scenarios from scheduling. Clean up with: bin/reset scenarios/cka/07-scheduling bin/doctor --fix