What is Node Selector in Kubernetes?

• Kubernetes uses labels on nodes
• Pod uses nodeSelector to match those labels
• Pod will only run on nodes that match labels

In Kubernetes, both nodes and pods use labels to control where workloads run. A node is first assigned a label such as disktype=ssd, which describes its characteristics. Then, when creating a pod, you specify a nodeSelector with the same key-value pair (disktype=ssd). The Kubernetes scheduler compares the pod’s selector with the available node labels. If a node has a matching label, the scheduler places (runs) the pod on that node. However, if no node in the cluster has the required label, Kubernetes cannot find a suitable node, so the pod remains in a Pending state until a matching node becomes available.

Step 1: Check available nodes

kubectl get nodes

Step 2: Add label to a specific node

kubectl label nodes node1 disktype=ssd

Verify label:

kubectl get nodes --show-labels

Step 3: Create Pod YAML with nodeSelector

vi pod.yaml

apiVersion: v1
kind: Pod
metadata:
  name: nginx-ssd
spec:
  containers:
  - name: nginx
    image: nginx
  nodeSelector:
    disktype: ssd

Node: Pod will only run on nodes that have “disktype: ssd”

Step 4: Apply Pod

kubectl apply -f pod.yaml

Step 5: Verify pod is running on correct node

kubectl get pods -o wide

If no node matches then pod status show as pending state. Kubernetes only schedules when labels match

Example:

Example (SSD node scheduling)

🎯 Scenario (Real-world)

You have a Kubernetes cluster with 2 nodes:

• node1 → high performance (SSD disk)
• node2 → normal storage (HDD)

You want your application (pod) to run only on the SSD node for better performance (like DB, analytics, etc.).

1. Check the available nodes
kubectl get nodes
2.  Label the node
kubectl label nodes node1 disktype=ssd
3. Verify the node label
kubectl get nodes --show-labels
4.  Create Pod YAML and used node selector in it
vi pod1.yaml
apiVersion: v1
kind: Pod
metadata:
  name: nginx-ssd
spec:
  containers:
  - name: nginx
    image: nginx
  nodeSelector:
    disktype: ssd
5. Deploy the POD
kubectl apply -f pod1.yaml
6. Verify where pod is running
kubectl get pods -o wide

Explanation

In this example, the node node1 was labeled with disktype=ssd, which describes its capability. When the pod was created, the nodeSelector field specified the same label (disktype=ssd). The Kubernetes scheduler compared the pod’s requirement with the available nodes and found that only node1 satisfied this condition. As a result, the pod was successfully scheduled and started on node1. If no node had the matching label, Kubernetes would not have been able to schedule the pod, and it would have remained in a Pending state until a suitable node became available