This ansible playbook is used to setup, configure and manage DCIC's JupyterHub. UMD's LDAP is used for authentication and every logged in user gets their own environment managed as a Docker container.
Jupyter's datascience-notebook is used which includes Python, R and Julia notebook environments with tons of useful third party libraries. But if other languages we can pick additional containers from docker-stacks or build our own. Leave an issue ticket if you'd like more things to be included.
First get this repository if you haven't already:
git clone https://github.com/umd-dcic/jupyterhub-ansible.git
cd jupyterhub-ansible
Edit hosts to point at the IP address of your server you want to target with
the playbook. Make sure the section name for your IP address matches the hosts
property in your jupyterhub.yml file. Edit the hub_ip property in your
group_vars/all/config.yml to point at the public IP address for your
JupyterHub instance.
We run JupyterHub under SSL. If you have a certificate (ssl.crt) and private key already (ssl.key) you can place them in this directory. Otherwise you can generate a self signed certificate by:
openssl req -x509 -newkey rsa:2048 -keyout ssl.key -out ssl.crt -days XXX
If there is a passphrase associated with your certificate you will want to:
cp group_vars/all/secret.template group_vars/all/secret.yml
and then add it the passphrase to group_vars/secret.yml.
Obviously you'll need to install Ansible in order to run the playbook. But once you've got Ansible you can install and start up JupyterHub by:
ansible-playbook jupyterhub.yml
If you want to work on this playbook you can use the supplied Vagrant setup which will bring up a virtual Ubuntu instance and configure that:
vagrant up