6. Kubernetes

6.1. Installation

To deploy to Kubernetes, we need the following components installed. On Linux, kubectl and minikube are single binary executable files.

• Install kubectl

  • kubectl documentation

• Install minikube

  • minikube documentation

• Install VirtualBox

6.2. Starting minikube cluster

Start minikube. This command effectively creates a virtual machine.

minikube start

You should see an output like the following.

😄  minikube v1.5.2 on Ubuntu 19.10
✨  Automatically selected the 'virtualbox' driver (alternates: [none])
💿  Downloading VM boot image ...
    > minikube-v1.5.1.iso.sha256: 65 B / 65 B [--------------] 100.00% ? p/s 0s
    > minikube-v1.5.1.iso: 143.76 MiB / 143.76 MiB [] 100.00% 127.66 MiB p/s 2s
🔥  Creating virtualbox VM (CPUs=2, Memory=2000MB, Disk=20000MB) ...
🐳  Preparing Kubernetes v1.16.2 on Docker '18.09.9' ...
💾  Downloading kubeadm v1.16.2
💾  Downloading kubelet v1.16.2
🚜  Pulling images ...
🚀  Launching Kubernetes ...
⌛  Waiting for: apiserver
🏄  Done! kubectl is now configured to use "minikube"

6.3. Quicktest

6.3.1. Deploy through an image

Type in the following to deploy a service through an image.

# create deployment
kubectl create deployment hello-minikube --image=k8s.gcr.io/echoserver:1.10

# expose deployment as a service
kubectl expose deployment hello-minikube --type=NodePort --port=8080

# get pod information
kubectl get pods

# get the service url
minikube service hello-minikube --url

# delete service
kubectl delete services hello-minikube

# delete deployment
kubectl delete deployment hello-minikube

6.3.2. Deploy through YAML configuration

Create a file called pod.yaml.

apiVersion: v1
kind: Pod
metadata:
  name: demo
spec:
  containers:
  - name: testpod
    image: alpine:3.5
    command: ["ping", "8.8.8.8"]

Then type in the following.

# create deployment
kubectl apply -f pod.yml

# get pod information
kubectl get pods

# get logs of pod
kubectl logs demo

## delete service
kubectl delete -f pod.yml

6.3.3. Pod creation

Create a file called pod.yaml.

apiVersion: v1
kind: Pod
metadata:
  name: student
spec:
  containers:
  - name: db
    image: db-app:local
    ports:
      - containerPort: 3306
    env:
      - name: MYSQL_ROOT_PASSWORD
        value: "oneoffcoder"
    volumeMounts:
      - name: db-scripts
        mountPath: /docker-entrypoint-initdb.d
  - name: rest
    image: rest-app:local
    ports:
      - containerPort: 5000
  volumes:
  - name: db-scripts
    hostPath:
      path: /hosthome/super/git/books/sphinx/docker/source/_static/code/compose/mysql/docker-entrypoint-initdb.d
      type: Directory

Now try running the following commands to create a pod and more.

# create deployment
kubectl apply -f student.yml

# get pod information
kubectl get pods
kubectl describe pod student

# get logs of pod
kubectl logs student db
kubectl logs student rest

# shell access
kubectl exec -it student --container db -- /bin/bash
kubectl exec -it student --container rest -- /bin/bash

## delete service
kubectl delete -f student.yml

6.3.4. Deployment creation

Build the docker images.

docker build --no-cache -t db-app:local .
docker build --no-cache -t rest-app:local .
docker build --no-cache -t ui-app:local .

Create student-deployment.yml with the following content.

apiVersion: apps/v1
kind: Deployment
metadata:
  name: student-deployment
  labels:
    app: student
spec:
  replicas: 1
  selector:
    matchLabels:
      app: student
  template:
    metadata:
      labels:
        app: student
    spec:
      containers:
      - name: db
        image: db-app:local
        ports:
          - containerPort: 3306
        env:
          - name: MYSQL_ROOT_PASSWORD
            value: "oneoffcoder"
        volumeMounts:
          - name: db-scripts
            mountPath: /docker-entrypoint-initdb.d
      - name: rest
        image: rest-app:local
        ports:
          - containerPort: 5000
      - name: ui
        image: ui-app:local
        ports:
          - containerPort: 80
      volumes:
      - name: db-scripts
        hostPath:
          path: /hosthome/super/git/books/sphinx/docker/source/_static/code/compose/mysql/docker-entrypoint-initdb.d
          type: Directory

Now try running the following commands to create a deployment and more.

# create deployment
kubectl create -f student-deployment.yml

# check deployment
kubectl get deployment

# check replication set
kubectl get rs

# get pod information
kubectl get pods
POD=student-deployment-75d56dc8f5-55rtt
kubectl describe pod $POD

# get logs
kubectl logs $POD db
kubectl logs $POD rest
kubectl logs $POD ui

# shell access
kubectl exec -it $POD --container db -- /bin/bash
kubectl exec -it $POD --container rest -- /bin/bash
kubectl exec -it $POD --container ui -- /bin/sh

# forward port
kubectl port-forward $POD 3306:3306
kubectl port-forward $POD 5000:5000
kubectl port-forward $POD 8080:80

## delete service
kubectl delete -f student-deployment.yml

6.4. Using local images

The following command enables us to reuse Minikube’s built-in docker daemon. This feature is useful to avoid building a Docker registry and pushing images into it. When you issue docker ps you will see the containers running on Minikube.

eval $(minikube docker-env)

6.5. Useful notes and commands

• minikube dashboard brings up the Kubernetes dashboard

• minikube ssh will SSH into the virutal machine

• kubectl get deployment gets the deployments in the cluster

• kubectl get events gets the events in the cluster

• kubectl get svc checks the services created

• /home (local) is mounted /hosthome (virtual machine)

• addons in ~/.minikube/addons (local) will be moved to the virtual machine on Minikube start/restart

6.6. Removing minikube cluster

Stop minikube. This command effectively halts the virtual machine.

minikube stop

Delete minikube. This command effectively deletes the virtual machine.

minikube delete