Skip to main content
NetApp Solutions

Use the NetApp DataOps Toolkit with JupyterHub

Contributors kevin-hoke
PDFs

The NetApp DataOps Toolkit for Kubernetes can be used in conjunction with JupyterHub. Using the NetApp DataOps Toolkit with JupyterHub enables end users to create volume snapshots for workspace backup and/or dataset-to-model traceability directly from within a Jupyter Notebook.

Initial Setup

Before you can use the DataOps Toolkit with JupyterHub, you must grant appropriate permissions to the Kubernetes service account that JupyterHub assigns to individual user Jupyter Notebook Server pods. JupyterHub uses the service account that is specified by the singleuser.serviceAccountName variable in your JupyterHub Helm chart configuration file.

Create Cluster Role for DataOps Toolkit

First, create a cluster role named 'netapp-dataops' that has the required Kubernetes API permissions for creating volume snapshots.

$ vi clusterrole-netapp-dataops-snapshots.yaml
---
kind: ClusterRole
apiVersion: rbac.authorization.k8s.io/v1
metadata:
  name: netapp-dataops-snapshots
rules:
- apiGroups: [""]
  resources: ["persistentvolumeclaims", "persistentvolumeclaims/status", "services"]
  verbs: ["get", "list"]
- apiGroups: ["snapshot.storage.k8s.io"]
  resources: ["volumesnapshots", "volumesnapshots/status", "volumesnapshotcontents", "volumesnapshotcontents/status"]
  verbs: ["get", "list", "create"]

$ kubectl create -f clusterrole-netapp-dataops-snapshots.yaml
clusterrole.rbac.authorization.k8s.io/netapp-dataops-snapshots created

Assign Cluster Role to Notebook Server Service Account

Create a role binding that assigns the 'netapp-dataops-snapshots' cluster role to the appropriate service account in the appropriate namespace. For example, if you installed JupyterHub in the 'jupyterhub' namespace, and you specified the 'default' service account via the singleuser.serviceAccountName variable, you would assign the the 'netapp-dataops-snapshots' cluster role to the 'default' service account in the 'jupyterhub' namespace as shown in the following example.

$ vi rolebinding-jupyterhub-netapp-dataops-snapshots.yaml
---
kind: RoleBinding
apiVersion: rbac.authorization.k8s.io/v1
metadata:
  name: jupyterhub-netapp-dataops-snapshots
  namespace: jupyterhub # Replace with you JupyterHub namespace
subjects:
- kind: ServiceAccount
  name: default # Replace with your JupyterHub singleuser.serviceAccountName
  namespace: jupyterhub # Replace with you JupyterHub namespace
roleRef:
  kind: ClusterRole
  name: netapp-dataops-snapshots
  apiGroup: rbac.authorization.k8s.io

$ kubectl create -f ./rolebinding-jupyterhub-netapp-dataops-snapshots.yaml
rolebinding.rbac.authorization.k8s.io/jupyterhub-netapp-dataops-snapshots created

Create Volume Snapshots Within Jupyter Notebook

Now, JupyterHub users can use the NetApp DataOps Toolkit to create volume snapshots directly from within a Jupyter Notebook as shown in the following example.

Figure showing input/output dialog or representing written content