DaemonSets for per-node workloads in Kubernetes - Commands & Configuration

Introduction

Sometimes you need to run a small program on every computer in your cluster. DaemonSets help you do this by making sure a copy of your program runs on each node automatically.

When you want to collect logs from every node in your cluster.

When you need to monitor the health of each node with a small agent.

When you want to run a network tool on every node to manage traffic.

When you need to apply security scanning software on all nodes.

When you want to run a storage driver on every node to handle disks.

Config File - daemonset.yaml

daemonset.yaml

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: node-agent
  labels:
    app: node-agent
spec:
  selector:
    matchLabels:
      app: node-agent
  template:
    metadata:
      labels:
        app: node-agent
    spec:
      containers:
      - name: node-agent
        image: busybox
        command: ["sh", "-c", "while true; do echo Hello from node $(hostname); sleep 30; done"]

This file creates a DaemonSet named node-agent. It runs a simple container on every node that prints a message every 30 seconds. The selector and template.metadata.labels ensure the DaemonSet manages the right pods. The command runs a loop printing the node's name.

Commands

This command creates the DaemonSet in your Kubernetes cluster, starting the pods on all nodes.

Terminal

kubectl apply -f daemonset.yaml

Expected Output

daemonset.apps/node-agent created

This command lists all DaemonSets in the current namespace to verify your DaemonSet is running.

Terminal

kubectl get daemonsets

Expected Output

NAME         DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR   AGE

node-agent   3         3         3       3            3           <none>          10s

This command shows the pods created by the DaemonSet and on which nodes they are running.

Terminal

kubectl get pods -l app=node-agent -o wide

Expected Output

NAME              READY   STATUS    RESTARTS   AGE   IP           NODE           NOMINATED NODE   READINESS GATES

node-agent-abc12   1/1     Running   0          15s   10.244.1.5   worker-node-1  <none>           <none>

node-agent-def34   1/1     Running   0          15s   10.244.2.5   worker-node-2  <none>           <none>

node-agent-ghi56   1/1     Running   0          15s   10.244.3.5   worker-node-3  <none>           <none>

→

-l app=node-agent - Filter pods by label to show only DaemonSet pods

→

-o wide - Show extra details including node names

This command shows detailed information about the DaemonSet, including events and pod status.

Terminal

kubectl describe daemonset node-agent

Expected Output

Name:           node-agent

Namespace:      default

Selector:       app=node-agent

Labels:         app=node-agent

Desired Number of Nodes Scheduled: 3

Current Number of Nodes Scheduled: 3

Number of Nodes Misscheduled: 0

Pods Status:    3 Running / 0 Waiting / 0 Succeeded / 0 Failed

Events:

  Type    Reason            Age   From                   Message

  ----    ------            ----  ----                   -------

  Normal  SuccessfulCreate  1m    daemonset-controller   Created pod: node-agent-abc12

  Normal  SuccessfulCreate  1m    daemonset-controller   Created pod: node-agent-def34

  Normal  SuccessfulCreate  1m    daemonset-controller   Created pod: node-agent-ghi56

Key Concept

A DaemonSet ensures one copy of a pod runs on every node automatically and stays running.

Common Mistakes

Not matching the selector labels with the pod template labels in the DaemonSet spec.

Kubernetes won't know which pods belong to the DaemonSet, so it won't manage them properly.

Make sure the selector.matchLabels exactly matches the labels in template.metadata.labels.

Applying the DaemonSet YAML without checking node readiness or taints.

Pods may not schedule on nodes that are not ready or have taints blocking them.

Verify nodes are ready and adjust tolerations in the DaemonSet spec if needed.

Using a container image that does not run or exits immediately.

Pods will crash or restart constantly, causing instability.

Use a container with a long-running process or a command that keeps running.

Summary

Create a DaemonSet YAML file that defines a pod to run on every node.

Apply the DaemonSet with kubectl apply to start pods on all nodes.

Use kubectl get daemonsets and kubectl get pods to verify pods are running on each node.

Use kubectl describe daemonset to see detailed status and events.