Schedule an application to a specified node with nodeSelector and node labels - Container Service for Kubernetes

Use node labels and a nodeSelector to direct pods to specific nodes in your ACK cluster. By default, the Kubernetes scheduler places pods automatically. Use nodeSelector when workloads have specific hardware requirements — for example, GPU-equipped nodes for machine learning tasks, SSD-backed nodes for I/O-intensive databases, or dedicated nodes for co-locating latency-sensitive services.

nodeSelector is the simplest scheduling constraint. Unlike taints and tolerations, it does not prevent other pods from running on the same node — it only controls where the pods it targets can land. For stronger isolation or more expressive rules, use node affinity or inter-pod affinity instead.

Prerequisites

Before you begin, ensure that you have:

An ACK managed cluster or an ACK dedicated cluster (ACK dedicated clusters can no longer be newly created)
A stateless workload (Deployment) or a stateful workload (StatefulSet) to deploy

Step 1: Label a node

Log on to the ACK console. In the left navigation pane, click Clusters.
On the Clusters page, click the name of your cluster. In the left navigation pane, choose Nodes > Nodes.
In the upper-right corner of the page, click Manage Labels and Taints. On the Labels tab, add a label to the target node. In this example, set Name to pod and Value to nginx. For more information, see Create and manage node labels.

Step 2: Deploy with a nodeSelector

Log on to the ACK console. In the left navigation pane, click Clusters.
On the Clusters page, click the name of your cluster. In the left navigation pane, choose Workloads > Deployments.

On the Deployments page, click Create from YAML and apply the following manifest. The nodeSelector field restricts the pods to nodes that carry the pod: nginx label.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nginx-deployment-basic
  labels:
    app: nginx
spec:
  replicas: 2
  selector:
    matchLabels:
      app: nginx
  template:
    metadata:
      labels:
        app: nginx
    spec:
      nodeSelector:
        pod: nginx      # Pods are scheduled only to nodes with this label.
      containers:
      - name: nginx
        image: anolis-registry.cn-zhangjiakou.cr.aliyuncs.com/openanolis/nginx:1.14.1-8.6
        ports:
        - containerPort: 80

On the Deployments page, click the name of the Deployment. On the Pods tab, check the node column for each pod. All pods should be running on nodes that carry the pod: nginx label. If a pod shows a different node, verify that the node label was applied correctly in Step 1.

What's next

For a full reference on nodeSelector and other node-assignment methods, see Assigning Pods to Nodes.
To control pod scheduling order and configure reverse scale-in during deployments and scale-out operations, see Customize elastic resource priority scheduling.