This repo contains the configuration files for the Jupyterhub cluster setup based on kubernetes, jupyterhub helm chart and fluxcd as gitops tool.
Our approach is to have a single jupyterhub cluster with user-dedicated IDEs in own pods for specific projects. The login in SSO via Azure Devops. Relevant credentials are deployed via Environment Variables on the Pods itself.
Go to: https://jupyter.analyticsinternal.iunera.com Log in with azure ad
- iunera Datascience environment
Open a Terminal in your Jupyter Notebook File->new->Terminal
Add the following command line
git config --global credential.helper store
Can be found jupyterhub.helm.yaml
- singleuser-secrets.yaml
apiVersion: v1
kind: Secret
metadata:
name: singleuser-secrets
namespace: jupyterhub
type: Opaque
stringData:
s3_access_key_id:
s3_access_key_secret:
s3_endpoint:
- custom-hub-secrets.yaml
apiVersion: v1
kind: Secret
metadata:
name: custom-hub-secrets
namespace: jupyterhub
type: Opaque
stringData:
hub.config.ConfigurableHTTPProxy.auth_token:
hub.config.JupyterHub.cookie_secret:
client_id:
client_secret:
tenant_id:
oauth_callback_url: https://<domain>/hub/oauth_callback
To provide more features and libs by default we have extended the jupyter/datascience-notebook container image in the Dockerfile
You might need some additional Libs installed. You can just do this inside your environment (File->new->Terminal) using pip or conda.
If you want to change the notebook images here e.g. the datascience-notebook create a branch and extend the Dockerfile. The Jupyterhub entry iunera Datascience environment LATEST-DEV IMAGE always pulls the latest image build from a branch.
If you are done send a Pull Request.
The new Stable Image needs to be applied.