This repository provides a reference deployment of JupyterHub, a multi-user Jupyter Notebook environment, on a single host using either Linux sudo processes or Docker using Vangrant and Ansible.
This setup provides a single-command way to startup start up a JupyterHub RHEL/CentOS machine. The only requirement is needed is vagrant. This setup heavily borrows from a number of ideas such as the JupyterHub deploy docker setup.
This repo contributes to what's already available in the following ways:
-
Provides the Vagrant+Ansible way to test a single-node JupyterHub installation on a clean CentOS VM.
-
Setup for CentoOS using a jupyterhub ansible role with four different spawner configurations available at provision-time:
- sudo spawner (jupyter notebook)
- sudo spawner (jupyter lab)
- docker spawner (jupyter lab, system user)
- image spawner (jupyter lab, system user, multiple images selector)
from the repo root dir type: vagrant up
The default provisioning runs docker image spawner, which allows the logged-in user to select a number of instances.
- role: ansible-role-centos-jupyterhub
jupyterhub_spawner: 'imagespawner'
jupyterhub_spawner_docker_image:
- 'natbusa/jupyterhub-scipy-lab'
- 'natbusa/jupyterhub-datascience-lab'
- 'natbusa/jupyterhub-tensorflow-lab'
- 'natbusa/jupyterhub-all-spark-lab'
generate_testusers: True
gen_test_username:
- amy
- joe
- sue
- bob
for more details about the JupyterHub ansible script, pleas refer to https://github.com/natbusa/ansible-role-centos-jupyterhub
On the host - check the system journal:
sudo journalctl -u jupyterhub
when running docker containers:
sudo docker logs <container-id>
list running and paused containers:
sudo docker ps -a