Summary - DaemonSet

Combine YAML and nodeSelector to write a DaemonSet.

Summary: DaemonSet

In this group:

A DaemonSet runs exactly one pod on each eligible node; when a node is added/removed, the pod is added/removed too.
YAML: apps/v1, kind: DaemonSet, selector + template; there is no replicas.
nodeSelector/nodeAffinity select nodes; tolerations allow tainted nodes.
Uses: log collection, monitoring, networking, storage and security agents; resource limits multiply with the node count.
Updates: RollingUpdate (maxUnavailable) or OnDelete; the rollout commands work.

kubectl get ds -A
kubectl rollout status ds/node-exporter -n izleme

Exercise A collector on every node

Write a DaemonSet template's spec section (starting with spec:): under tolerations, key: node-role.kubernetes.io/control-plane, operator: Exists, effect: NoSchedule; container named fluent-bit, image fluent/fluent-bit:3.1.

Show example solution

spec:
  tolerations:
    - key: node-role.kubernetes.io/control-plane
      operator: Exists
      effect: NoSchedule
  containers:
    - name: fluent-bit
      image: fluent/fluent-bit:3.1

Test yourself

If a log-collector DaemonSet doesn't run on the control-plane node, what is the likely cause?
- The image is wrong
- The namespace is wrong
- There's no toleration for the NoSchedule taint on the node
- replicas is missing
Show answer

There's no toleration for the NoSchedule taint on the node — Control-plane nodes are tainted by default.

Allow the taint.

      ____:
        - key: node-role.kubernetes.io/control-plane
          operator: Exists
          effect: NoSchedule

Show answer

      tolerations:
        - key: node-role.kubernetes.io/control-plane
          operator: Exists
          effect: NoSchedule

Match.

DaemonSet

Per node
nodeSelector

Node selection
tolerations

Taint allowance
OnDelete

Controlled update

Next lesson: What is a Job?

We Are Enhancing Your Experience

Cookie Preferences

Strictly Necessary Cookies

Analytical Cookies