MicroK8s

Installing

MicroK8s

In order to run this project you must first install MicroK8s in your host machine. You can install it by following the instructions at the MicroK8s official web site according to your operating system.

In this tutorial we're running MicroK8s in a MacOS. So take everything with a grain of salt and add the necessary changes for your OS.

Enabling addons

After installing MicroK8s, you must enable the necessary addons for this project to work. In this case, only the ingress, which you can enable by running

microk8s enable ingress

[OPTIONAL] Enabling Kubernetes dashboard

Yet To be written

Running the app

1. Building and pushing Docker image

In order to run the express server in your MicrK8s/Kubernetes cluster you must first build and push a docker image of the project to Docker Hub. You can find instructions of how to do this in Docker Hub tutorials page. It is necessary to have some knowledge on Docker.

In general, after creating a Docker Hub account, you must use Docker to build the current image described in the Dockerfile and then push it to Docker Hub after loging in. It would be something like this:

docker login
docker build -t express-app .
docker push express-app

2. Changing MicroK8s to use the pushed image

Currently MicroK8s is configured to use my pushed Docker image (gotassinari/express-microk8s). You must changed it for your pushed image in deployment.yml.

For example:

public class Hello1
apiVersion: apps/v1
kind: Deployment
metadata:
  name: express
spec:
  replicas: 3
  selector:
    matchLabels:
      app: express
  template:
    metadata:
      labels: # labels to select/identify the deployment
        app: express
    spec:     # pod spec                  
      containers: 
      - name: express 
-        image: gotassinari/express-microk8s
+        image: yourdockerhubaccount/express-microk8s
        ports:
        - containerPort: 3000

3. Running your app in MicroK8s cluster

Now you're all set to run the application in your cluster. As MicroK8s only runs in Linux OS, MacOS (and Windows, I suppose) actually run it in a Virtual Machine.

In MacOS you can see this virtual machine with multipass (which is needed) to run MicroK8s in MacOS). Just run multipass list and you should see a running microk8s-vm.

$ multipass list
                                                                           
Name                    State             IPv4             Image
microk8s-vm             Running           192.168.64.2     Ubuntu 18.04 LTS
                                          10.1.254.64

As MicroK8s is running in a VM you must know that everytime you run a MicroK8 command in your host machine it is actually running it in microk8s-vm and showing its output in your console.

Now, in order to run the app, you must create the deployment defined in deployment.yml. Just execute the following command:

microk8s kubectl create -f ./deployment.yml

Now you should have your app running with 3 replicas. You can observe that running

$ microk8s kubectl get deployment

NAME                  READY   UP-TO-DATE   AVAILABLE   AGE
express               3/3     3            3           2s

and

microk8s kubectl get pods

NAME                                   READY   STATUS    RESTARTS       AGE
express-8c84cd9fd-sxp7k                1/1     Running   0              41s
express-8c84cd9fd-w6pcs                1/1     Running   0              41s
express-8c84cd9fd-krhx5                1/1     Running   0              41s

This shows that you have a deployment running with 3 pods/replicas.

4. Exposing the pods to your host machine under a service

Once you have your pods running, you must now expose it to your host machine using a service and ingress. This will allow you to make requests to the express server from you host machine.

It is actually quite simple. First, you must create the service, which will act as a load balancer for the pods inside your MicroK8s cluster. To do that, run

microk8s kubectl create -f ./service.yml

You should be able to see your service is running using the following command

$ microk8s kubectl get service

NAME         TYPE           CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
kubernetes   ClusterIP      10.152.183.1     <none>        443/TCP          17d
express      LoadBalancer   10.152.183.247   <pending>     3000:31349/TCP   22s

Now, you must expose this service to your host machine using an ingress. Again, it is quite straightforward. Just run

microk8s kubectl create -f ./ingress.yml

You can see your ingress is running using the following command

$ microk8s kubectl get ingress 

NAME              CLASS    HOSTS   ADDRESS     PORTS   AGE
minimal-ingress   public   *       127.0.0.1   80      25s

5. Accessing your app from your host machine

Now you must only find the node IP in your host machine and make a request to the express app. To to this, execute the following command

$ microk8s kubectl get node -o wide

NAME          STATUS   ROLES    AGE   VERSION                    INTERNAL-IP    EXTERNAL-IP   OS-IMAGE             KERNEL-VERSION       CONTAINER-RUNTIME
microk8s-vm   Ready    <none>   17d   v1.22.3-3+9ec7c40ec93c73   192.168.64.2   <none>        Ubuntu 18.04.6 LTS   4.15.0-161-generic   containerd://1.5.2

The node IP will be under INTERNAL-IP. Now, all you ahve to do is access this IP in your web browser. It should show a hello-world page just like this one:

And there it goes! Everything is up and running!