How to quickly setup Kubernetes under Ubuntu.

Required OS:

• Ubuntu 20 LTS Tested:
• Ubuntu 20.04.3 LTS

TIP: See:

• https://gitlab.com/hpaluch/mysql-micro-k8s for example how to setup WordPress+MySQL with OpenEBS replicated volumes
• https://gitlab.com/hpaluch/openebs-to-pdf how to print OpenEBS site documentation to PDF

At first you have to install snapd packages using: (see a bit unhelpful https://snapcraft.io/docs/installing-snap-on-ubuntu)

sudo apt-get install snapd

WARNING! There also exists package snap which has nothing common with snapcraft (it is some kind of DNA tool :-)

Now we have to follow guide from:

• https://microk8s.io/

sudo snap install microk8s --classic
microk8s status --wait-ready

You will likely get permission error and will be guided how to fix it:

sudo usermod -a -G microk8s $USER
sudo chown -f -R $USER ~/.kube

Now relogin (you can use newgrp microk8s instead, but it will invoke new shell)

After relogin try:

microk8s status --wait-ready
# we changed plugint list a bit
microk8s enable dashboard dns rbac

Normally you have to run kubectl using microk8s wrapper: (from guide)

microk8s kubectl get all --all-namespaces

NOTE: you can ignore the messages from dashboard (about RBAC). To get token you just need to run:

microk8s dashboard-proxy

It will print Dashboard URL and necessary token

What next - you can follow for example:

• https://microk8s.io/docs

Example from above guide:

microk8s kubectl create deployment nginx --image=nginx

Now you have to wait until command below will show Status: Running and READY: 1/:

microk8s kubectl get pods
NAME                     READY   STATUS    RESTARTS   AGE
nginx-6799fc88d8-kc7zq   1/1     Running   0          59s

How to reach Nginx Pod? According to official docs

• https://v1-19.docs.kubernetes.io/docs/concepts/services-networking/service/#publishing-services-service-types
• There many ways, we will use
• NodePort So we can follow:
• https://kubernetes.io/docs/tasks/access-application-cluster/service-access-application-cluster/ Do this:
• get deployment name:

   microk8s kubectl get deployments 
   NAME    READY   UP-TO-DATE   AVAILABLE   AGE
   nginx   1/1     1            1           30m

• so we have verified that our deployment is called nginx
• we can expose it via service nginx-svc:

  kubectl expose deployment nginx \
    --type=NodePort --name=nginx-svc --port=80

• now you have to list what port was exposed using:

  microk8s kubectl get svc nginx-svc
  NAME        TYPE       CLUSTER-IP      EXTERNAL-IP   PORT(S)        AGE
  nginx-svc   NodePort   10.152.183.88   <none>        80:32001/TCP   109s

• in our case you have to use exposed port 32001, for example:

  curl http://127.0.0.1:32001

• above port 32001 should be reachable also from any remote host using url like http://IP_OF_YOUR_K8S:32001 (ensure that your firewall is not blocking it)

Once you are finished with k8s you should stop it to relieve your computer resources:

microk8s stop

• Use Ingress plugin to expose services
  • https://microk8s.io/docs/addon-ingress
• Add node to cluster:
  • https://microk8s.io/docs/clustering
• Learn from Google's k8s examples:
  • https://github.com/GoogleCloudPlatform/kubernetes-engine-samples
• Learn Helm:
  • https://docs.bitnami.com/tutorials/create-your-first-helm-chart/
• Learn Helm from Bitnami charts:
  • https://github.com/bitnami/charts

Resources:

• https://virtualizationreview.com/articles/2019/02/01/microk8s-part-3-how-to-deploy-a-pod-in-kubernetes.aspx