Home/

...

/

Self-Heal from a Pod Failure

Self-Heal from a Pod Failure

Learn how Kubernetes self-heals from a Pod failure.

In this lesson, we’ll use a Kubernetes Deployment to deploy five replicas of a Pod. After that, we’ll manually delete a Pod and see how Kubernetes self-heals.

We’ll use the deploy.yml file in the /usercode directory. It defines five replicas of the app we containerized in previous chapters.

kind: Deployment       #   <<== Type of object being defined
apiVersion: apps/v1    #   <<== Version of object to deploy
metadata:
  name: qsk-deploy
spec:
  replicas: 5                #   <<== How many Pod replicas
  selector:
    matchLabels:            # <<== Tells the Deployment controller
      project: qsk-book     #  <<== to manage Pods with this label
  template:
    metadata:
      labels:
        project: qsk-book   #  <<== Give all replicas this label
    spec: 
      containers:
      - name: qsk-pod
        imagePullPolicy: Always   # <<== Never use images from local machine
        ports:
        - containerPort: 8080             # <<== Network port
        image: nigelpoulton/qsk-book:1.0  # <<== Container image to use
Playground

We use the terms Pod, instance, and replica to mean the same ...