Deploy Kubernetes and JupyterHub on Jetstream with Magnum


June 14, 2019

This tutorial is obsolete, please checkout the updated version at

This tutorial deploys Kubernetes on Jetstream with Magnum and then JupyterHub on top of that using zero-to-jupyterhub.

In my previous tutorials I deployed Kubernetes using Kubespray. The main driver to using Magnum is that there is support for autoscaling, i.e. create and destroy Openstack instances based on the load on JupyterHub. I haven’t tested that yet, though, that will come in a following tutorial.

Magnum is a technology built into Openstack to deploy Container Orchestration engines based on templates. The main difference with kubespray is that is way less configurable, the user does not have access to modify those templates but has just a number of parameters to set. Instead Kubespray is based on ansible and the user has full control of how the system is setup, it also supports having more High Availability features like multiple master nodes. On the other hand, the ansible recipe takes a very long time to run, ~30 min, while Magnum creates a cluster in about 10 minutes.

Setup access to the Jetstream API

First install the OpenStack client, please use these exact versions, also please run at Indiana, which currently has the Rocky release of Openstack, the TACC deployment has an older release of Openstack.

pip install python-openstackclient==3.16 python-magnumclient==2.10

Load your API credentials from, check documentation of the Jetstream wiki for details.

You need to have a keypair uploaded to Openstack, this just needs to be done once per account. See the Jetstream documentation under the section “Upload SSH key - do this once”.

Create the cluster with Magnum

As usual, checkout the repository with all the configuration files on the machine you will use the Jetstream API from, typically your laptop.

git clone
cd jupyterhub-deploy-kubernetes-jetstream
cd kubernetes_magnum

Now we are ready to use Magnum to first create a cluster template and then the actual cluster, edit first and set the parameters of the cluster on the top. Also make sure to set the keypair name. Finally run:


I have setup a test cluster with only 1 master node and 1 normal node but you can modify that later.

Check the status of your cluster, after about 10 minutes, it should be in state CREATE_COMPLETE:

openstack coe cluster show k8s

Configure kubectl locally

Install the kubectl client locally, first check the version of the master node:

openstack server list # find the floating public IP of the master node (starts with 149_
ssh fedora@$IP
kubectl version

Now install the same version following the Kubernetes documentation

Now configure kubectl on your laptop to connect to the Kubernetes cluster created with Magnum:

mkdir kubectl_secret
cd kubectl_secret
openstack coe cluster config k8s

This downloads a configuration file and the required certificates.

and returns export KUBECONFIG=/absolute/path/to/config

See also the script to automate this step, but it requires to already have setup the environment variable.

execute that and then:

kubectl get nodes

Configure storage

Magnum configures a provider that knows how to create Kubernetes volumes using Openstack Cinder, but does not configure a storageclass, we can do that with:

kubectl create -f storageclass.yaml

We can test this by creating a Persistent Volume Claim:

kubectl create -f persistent_volume_claim.yaml

kubectl describe pv

kubectl describe pvc
Name:            pvc-e8b93455-898b-11e9-a37c-fa163efb4609
Annotations: cinder-dynamic-provisioner
Finalizers:      []
StorageClass:    standard
Status:          Bound
Claim:           default/pvc-test
Reclaim Policy:  Delete
Access Modes:    RWO
Capacity:        5Gi
Node Affinity:   <none>
    Type:       Cinder (a Persistent Disk resource in OpenStack)
    VolumeID:   2795724b-ef11-4053-9922-d854107c731f
    ReadOnly:   false
    SecretRef:  nil
Events:         <none>

We can also test creating an actual pod with a persistent volume and check that the volume is successfully mounted and the pod started:

kubectl create -f ../alpine-persistent-volume.yaml
kubectl describe pod alpine

Note about availability zones

By default Openstack servers and Openstack volumes are created in different availability zones. This created an issue with the default Magnum templates because we need to modify the Kubernetes scheduler policy to allow this. Kubespray does this by default, so I created a fix to be applied to the Jetstream Magnum templates, this needs to be re-applied after every Openstack upgrade.

Install Helm

The Kubernetes deployment from Magnum is not as complete as the one out of Kubespray, we need to setup helm and the NGINX ingress ourselves. We would also need to setup a system to automatically deploy HTTPS certificates, I’ll add this later on.

First install the Helm client on your laptop, make sure you have configured kubectl correctly.

Then we need to create a service account to give enough privilege to Helm to reconfigure the cluster:

kubectl create -f tiller_service_account.yaml

Then we can create the tiller pod inside Kubernetes:

helm init --service-account tiller --wait --history-max 200
kubectl get pods --all-namespaces
NAMESPACE     NAME                                       READY   STATUS    RESTARTS   AGE
kube-system   coredns-78df4bf8ff-f2xvs                   1/1     Running   0          2d
kube-system   coredns-78df4bf8ff-pnj7g                   1/1     Running   0          2d
kube-system   heapster-74f98f6489-xsw52                  1/1     Running   0          2d
kube-system   kube-dns-autoscaler-986c49747-2m64g        1/1     Running   0          2d
kube-system   kubernetes-dashboard-54cb7b5997-c2vwx      1/1     Running   0          2d
kube-system   openstack-cloud-controller-manager-tf5mc   1/1     Running   3          2d
kube-system   tiller-deploy-6b5cd64488-4fkff             1/1     Running   0          20s

And check that all the versions agree:

helm version
Client: &version.Version{SemVer:"v2.11.0", GitCommit:"2e55dbe1fdb5fdb96b75ff144a339489417b146b", GitTreeState:"clean"}
Server: &version.Version{SemVer:"v2.11.0", GitCommit:"2e55dbe1fdb5fdb96b75ff144a339489417b146b", GitTreeState:"clean"}

Setup NGINX ingress

We need to have the NGINX web server to act as front-end to the services running inside the Kubernetes cluster.

Open HTTP and HTTPS ports

First we need to open the HTTP and HTTPS ports on the master node, you can either connect to the Horizon interface, create new rule named http_https, then add 2 rules, in the Rule drop down choose HTTP and HTTPS; or from the command line:

openstack security group create http_https
openstack security group rule create --ingress --protocol tcp --dst-port 80 http_https 
openstack security group rule create --ingress --protocol tcp --dst-port 443 http_https 

Then you can find the name of the master node in openstack server list then add this security group to that instance:

openstack server add security group  k8s-xxxxxxxxxxxx-master-0 http_https

Install NGINX ingress with Helm


Note, the documentation says we should add this annotation to ingress with kubectl edit ingress -n jhub, but I found out it is not necessary:

  annotations: nginx

If this is correctly working, you should be able to run curl localhost from the master node and get a Default backend: 404 message.

Install JupyterHub

Finally, we can go back to the root of the repository and install JupyterHub, first create the secrets file:


Then edit secrets.yaml and modify the hostname under hosts to display the hostname of your master Jetstream instance, i.e. if your instance public floating IP is, the hostname should be (without http://).

You should also check that connecting with your browser to shows default backend - 404, this means NGINX is also reachable from the internet, i.e. the web port is open on the master node.



Connect with your browser to to check if it works.

Issues and feedback

Please open an issue on the repository to report any issue or give feedback. Also you find out there there what I am working on next.


Many thanks to Jeremy Fischer and Mike Lowe for solving all my tickets, this required a lot of work on their end to make it working.