DaemonSets

DaemonSets — One Pod Per Node

A DaemonSet ensures that exactly one copy of a pod runs on every node in the cluster (or on specified nodes). When a new node is added, the pod is automatically scheduled there.

DaemonSets are perfect for infrastructure workloads like logging agents, monitoring exporters, and network plugins.

How DaemonSets Work

  Cluster with 3 nodes:
  ┌──────────┐    ┌──────────┐    ┌──────────┐
  │  Node 1  │    │  Node 2  │    │  Node 3  │
  │ ┌──────┐ │    │ ┌──────┐ │    │ ┌──────┐ │
  │ │Pod A │ │    │ │Pod A │ │    │ │Pod A │ │
  │ └──────┘ │    │ └──────┘ │    │ └──────┘ │
  └──────────┘    └──────────┘    └──────────┘

  New node joins:
  ┌──────────┐    ┌──────────┐    ┌──────────┐    ┌──────────┐
  │  Node 1  │    │  Node 2  │    │  Node 3  │    │  Node 4  │
  │ ┌──────┐ │    │ ┌──────┐ │    │ ┌──────┐ │    │ ┌──────┐ │
  │ │Pod A │ │    │ │Pod A │ │    │ │Pod A │ │    │ │Pod A │ │
  │ └──────┘ │    │ └──────┘ │    │ └──────┘ │    │ └──────┘ │
  └──────────┘    └──────────┘    └──────────┘    └──────────┘
  Automatically deployed to Node 4!

Creating a DaemonSet

apiVersion: apps/v1
kind: DaemonSet
metadata:
  name: fluentd
  namespace: kube-system
spec:
  selector:
    matchLabels:
      name: fluentd
  template:
    metadata:
      labels:
        name: fluentd
    spec:
      tolerations:
        - key: node-role.kubernetes.io/control-plane
          effect: NoSchedule
      containers:
        - name: fluentd
          image: fluentd:v1.16
          resources:
            limits:
              memory: "200Mi"
              cpu: "100m"
          volumeMounts:
            - name: varlog
              mountPath: /var/log
      volumes:
        - name: varlog
          hostPath:
            path: /var/log

tolerations — The control plane node has a taint that prevents normal pods from running. This DaemonSet tolerates that taint so it runs everywhere, including the master.

Common DaemonSet Use Cases

Logging — Collect logs from every node (Fluentd, Filebeat).

Monitoring — Export node metrics (Prometheus Node Exporter).

Networking — Manage network plugins (Calico, Weave Net).

Storage — Provide storage drivers on every node (Ceph, GlusterFS).

Managing DaemonSets

# Apply a DaemonSet
kubectl apply -f daemonset.yaml

# Check which pods are running on which nodes
kubectl get pods -o wide

# Update the DaemonSet image
kubectl set image daemonset/fluentd fluentd=fluentd:v1.17

# Rollback
kubectl rollout undo daemonset/fluentd