How can you restrict a pod to a specific set of nodes in Kubernetes?

Using a nodeSelector is the correct approach to restrict a pod to a specific set of nodes in Kubernetes. A nodeSelector allows you to define specific labels that a node must have for the pod to be scheduled onto it. When you declare a nodeSelector in the pod's specification, Kubernetes will only place that pod on nodes that match the specified labels.

For example, if you have nodes labeled with specific roles or attributes (like role=frontend), you can configure the pod to be scheduled only on those nodes by adding the nodeSelector to its configuration. This provides a very straightforward way to control where your pods run based on the characteristics of the nodes.

The other options do not provide a mechanism for restricting pods to specific nodes. For instance, a nodeLimit is not a standard Kubernetes feature for pod scheduling, and specifying the node’s IP address does not work because Kubernetes scheduling relies on labels rather than fixed IPs for scalable, dynamic environments. Resource quotas are utilized to limit the total resources consumed by a namespace but do not dictate on which nodes the pods may run.