Kubernetes - Scheduling
If a pod has a preferred podAntiAffinity rule to avoid nodes with pods labeled
app: cache, what is the expected behavior when all nodes have such pods?