Using node selectors and node name

Node selectors are a very simple type of placement control in Kubernetes. Each Kubernetes node can be labeled with one or more labels in the metadata block, and Pods can specify a node selector.

To label an existing node, you can use the kubectl label command:

> kubectl label nodes node1 cpu_speed=fast

In this example, we're labeling our node1 node with the label cpu_speed and the value fast.

Now, let's assume that we have an application that really needs fast CPU cycles to perform effectively. We can add a nodeSelector to our workload to ensure that it is only scheduled on nodes with our fast CPU speed label, as shown in the following code snippet:

pod-with-node-selector.yaml

apiVersion: v1
kind: Pod
metadata:
  name: speedy-app
spec:
  containers:
  - name: speedy-app
    image: speedy-app:latest
    imagePullPolicy: IfNotPresent
  nodeSelector...