MicroK8s on Google Cloud Platform

Daz Wilkin

Published in

Google Cloud - Community

3 min readJun 21, 2019

https://microk8s.io/

How cool is this!?

Setup

PROJECT=[[YOUR-PROJECT]]
INSTANCE=[[YOUR-INSTANCE]] # Perhaps 'microk8s'
ZONE=[[YOUR-ZONE]]gcloud compute instances create ${INSTANCE} \
--machine-type=custom-2-8192 \
--preemptible \
--tags=microk8s \
--image-family=ubuntu-minimal-1904 \
--image-project=ubuntu-os-cloud \
--zone=${ZONE} \
--project=${PROJECT} \
--metadata=startup-script='
!# /bin/bash
sudo snap install microk8s --classic --channel=1.15/stable
'

NB Use whichever Ubuntu version you prefer. I really wanted Ubuntu Core but I think -minimial- is likely as good as. Ubuntu comes with snap preinstalled.

If you want to save time prefixing microk8s.kubectl commands with sudo, you may add your user to the instance’s microk8s group with:

gcloud compute ssh ${INSTANCE} \
--zone=${ZONE} \
--project=${PROJECT} \
--command="sudo usermod -a -G microk8s $(whoami)"

Then:

gcloud compute ssh ${INSTANCE} \
--zone=${ZONE} \
--project=${PROJECT}

Kubernetes

microk8s.kubectl cluster-info
Kubernetes master is running at https://127.0.0.1:16443
Heapster is running at ...
Grafana is running at ...
InfluxDB is running at ...microk8s.kubectl get nodes
NAME       STATUS   ROLES    AGE   VERSION
microk8s   Ready    <none>    5m   v1.15.0

Then:

microk8s.kubectl run nginx --image=docker.io/nginx --port=80microk8s.kubectl get pods
NAME                     READY   STATUS    RESTARTS   AGE
nginx-5d756c54b9-hm97d   1/1     Running   0          48smicrok8s.kubectl scale deployment/nginx --replicas=3microk8s.kubectl get pods
NAME                     READY   STATUS    RESTARTS   AGE
nginx-5d756c54b9-gh5m6   1/1     Running   0          7s
nginx-5d756c54b9-hm97d   1/1     Running   0          71s
nginx-5d756c54b9-j869t   1/1     Running   0          7smicrok8s.kubectl expose deployment/nginx --type=NodePortmicrok8s.kubectl describe service/nginx
Name:                     nginx
Namespace:                default
Labels:                   run=nginx
Annotations:              <none>
Selector:                 run=nginx
Type:                     NodePort
IP:                       10.152.183.77
Port:                     <unset>  80/TCP
TargetPort:               80/TCP
NodePort:                 <unset>  32592/TCP
Endpoints:                10.1.1.10:80,10.1.1.11:80,10.1.1.12:80
Session Affinity:         None
External Traffic Policy:  Cluster
Events:                   <none>

And, of course:

# Programmatically grab the NodePort value (#32592 above)
PORT=$(\
  microk8s.kubectl get service/nginx \
  --output=jsonpath="{.spec.ports[0].nodePort}")curl http://localhost:${PORT}

Returns:

<!DOCTYPE html>
<html>
<head>
<title>Welcome to nginx!</title>
<style>
    body {
        width: 35em;
        margin: 0 auto;
        font-family: Tahoma, Verdana, Arial, sans-serif;
    }
</style>
</head>
<body>
<h1>Welcome to nginx!</h1>
<p>If you see this page, the nginx web server is successfully installed and
working. Further configuration is required.</p><p>For online documentation and support please refer to
<a href="http://nginx.org/">nginx.org</a>.<br/>
Commercial support is available at
<a href="http://nginx.com/">nginx.com</a>.</p><p><em>Thank you for using nginx.</em></p>
</body>
</html>

Or, if you’d prefer:

gcloud compute ssh ${INSTANCE} \
--ssh-flag="-L ${PORT}:localhost:${PORT}" \
--zone=${ZONE} \
--project=${PROJECT}

NB Because you must run the gcloud command from your local workstation, you won’t have ${PORT} set. So you must replace ${PORT} with the value (32592 in my case) or set the variable’s value locally.
NB There’s no requirement that the remote ${PORT} value be mapped to same port on your workstation; you could also use 80:localhost:${PORT} to access the Nginx service on :80.

And then you can access the service from your local workstation:

google-chrome http://localhost:${PORT}

Conclusion

As a Googler, I’m spoiled by Kubernetes Engine *but* sometimes you just want the simplest possible cluster *or* you want to develop|deploy locally. MicroK8s is excellent.

Follow-up

Securely accessing the remote cluster locally.

The API server is available on :16443. If we can create a local kubernetes config file *and* we can port-forward to :16443, then we can access the cluster remotely:

gcloud compute ssh ${INSTANCE} \
--zone=${ZONE} \
--project=${PROJECT} \
--ssh-flag="-L 16443:localhost:16443"

Then, from another terminal (while the above ssh port-forward is running), we can copy the microk8s config file (

gcloud compute scp \
${INSTANCE}:/var/snap/microk8s/current/credentials/client.config \
${PWD} \
--zone=${ZONE} \
--project=${PROJECT}

And then, either use KUBECONFIG:

KUBECONFIG=${PWD}/client.config
kubectl get pods

Or, explicitly referencing the config on each command:

kubectl --kubeconfig=${PWD}/client.config get pods

For completeness, Google’s Cloud SDK (gcloud) andkubectl are both available as Snaps if your local workstation also uses Snap.

MicroK8s on Google Cloud Platform

Setup

Kubernetes

Conclusion

Follow-up

Written by Daz Wilkin