Minikube Deployment

How to deploy the ScienceMesh IOP using Minikube?

This tutorial demonstrates how to deploy both REVA and Phoenix on a local Minikube cluster with a single helm chart.


Launch minikube and configure the helm-tiller add-on

minikube start
minikube addons enable helm-tiller

Note: If it is the first time you’re running minikube, take a look to the different VM drivers available for the `–driver’ flag to pick the best one for your system.

Deploy REVA and Phoenix using helm charts

You can find more details on how this example works in the REVA tutorials. For the time being, you can install an example chart for ownCloud Phoenix by cloning this repo and running:

git clone

helm install phoenix sciencemesh/k8s/phoenix

Once this is done, you can install and plug-in the IOP by fetching the sciencemesh charts from the official helm repo and overriding some revad settings:

helm repo add sciencemesh

helm install iop \
  --set revad.workingDir=/go/src/github/cs3org/reva/examples/oc-phoenix \
  --set revad.args="{-dev-dir,.}" \

Afterwards, you can verify the deployments and their readyness by running:

kubectl get deployments -o wide
NAME                 READY   UP-TO-DATE   AVAILABLE   AGE     CONTAINERS          IMAGES                SELECTOR
iop-revad            1/1     1            1           50s     revad               cs3org/revad:v0.1.0,
phoenix-deployment   1/1     1            1           57s     phoenix-container   owncloud/phoenix      app=phoenix

Forward ports

In order to access the Phoenix web UI from your browser, you’ll need to create some port-forwards to the services runninng in the cluster:

kubectl port-forward svc/iop-revad 20080:20080 & \
kubectl port-forward svc/phoenix-svc 8300:8300 &

Now just browse to http://localhost:8300 to verify everything is up and running. Try to log in by using username einstein and password relativity.


Once done testing, you can quickly tear down the deployment by running:

helm delete iop
helm delete phoenix
# If you want to remove the local cluster as well:
minikube delete