Making your Raspberry Pi works for your MicroK8s cluster

MicroK8s Cluster on Raspberry Pi 4 Model B 8GB (Part I)

(Total Setup Time: 40 mins)

I am very excited to receive my Raspberry Pi 4 today. Since I do not have the microHDMI cable, I decided to go for a headless install. With this new Pi 8GB, I plan to check out on the MicroK8s , a lightweight upstream K8s. This tutorial shows my steps for setting up the MicroK8s Cluster on Raspberry Pi 4 Model B 8GB.

Raspberry Pi OS Preparation

(15 mins)

Due to the 64-bit requirements for MicroK8s, I am using Ubuntu Server (64-bit) as my OS. Since this is going to be a headless install, after burning the image onto my 32GB SD card, I created an empty file named ssh at d:/boot (you may notice that I am working on a Windows PC).

Powering Up

(10 mins)

After powering the Raspberry Pi up with the USB-C power adapter, it can be accessed via PuTTy :

putty-raspberry-pi-8gb

The default username is ubuntu and password is ubuntu, be sure to change password as always. Update the OS to the latest by:

sudo apt update
sudo apt upgrade

Change its hostname by running:

sudo nano /etc/hostname
sudo nano /etc/hosts

Enable memory cgroup, by opening the file:

sudo nano /boot/firmware/cmdline.txt

Next, add the following at the end of the line:

cgroup_enable=memory cgroup_memory=1

Reboot after the above modification and check that memory group is enabled (“1”):

grep mem /proc/cgroups | awk '{ print $4 }'

Installing MicroK8s on Raspberry Pi

(10 mins)

After the reboot, install microK8s:

sudo snap install microk8s --classic

Add the ubuntu user to the microk8s group:

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

Ensure that the ubuntu node is ready (without setting the cgroup memory, it will always report as not ready):

su - ubuntu
microk8s kubectl get nodes

Using MicroK8s

(5 mins)

For checking MicroK8s status and ensure that all services are up:

microk8s status --wait-ready

If you wish to turn on any extra services that you want (microk8s enable –help for more options), for example enabling dashboard, dns and registry:

microk8s enable dashboard dns registry

In order to access dashboard without a token, add –enable-skip-login option to the spec section (microk8s edit uses the vim editor ):

sudo microk8s kubectl -n kube-system edit deploy kubernetes-dashboard -o yaml
spec:
  progressDeadlineSeconds: 600
  replicas: 1
  revisionHistoryLimit: 10
  selector:
    matchLabels:
      k8s-app: kubernetes-dashboard
  strategy:
    rollingUpdate:
      maxSurge: 25%
      maxUnavailable: 25%
    type: RollingUpdate
  template:
    metadata:
      creationTimestamp: null
      labels:
        k8s-app: kubernetes-dashboard
    spec:
      containers:
      - args:
        - --auto-generate-certificates
        - --namespace=kube-system
        - --enable-skip-login

Next, to serve the dashboard in the background (the skip button will appear only after a reboot or by running microk8s stop, followed by microk8s start):

microk8s kubectl proxy --accept-hosts=.\* --address=0.0.0.0 &

Finally, for viewing dashboard from my windows laptop (the grey text below denotes IP of your Pi). Hit the skip button to view the dashboard:

http://192.168.xx.xx:8001/api/v1/namespaces/kube-system/services/https:kubernetes-dashboard:/proxy
kubernetes

To add leaf nodes to the MicroK8s Cluster on Raspberry Pi 4 Model B 8GB, please check this out.

Troubleshooting

microk8s is not available on stable for this architecture (armhf)

Since MicroK8s is only available on 64-bit architecture (amd64, arm64, ppc64el), I have switched to using 64-bit OS instead.

Finding IP address of the Pi

Mac addresses are assigned by device manufacturers and you can find those from here . To find out the IP, you can use the following command:

# You may try e4-5f-01 too
arp -a | findstr dc-a6-32

Depending on your network configuration, if you are still unable to figure out the IP, you can use nmap to scan your network (replacing the IP accordingly):

nmap -sn 192.168.xx.0/24

Below is the sample nmap output:

MAC Address: DC:A6:32:xx:xx:xx (Raspberry Pi Trading)
Nmap scan report for 192.168.xx.xx