Jupyterhub backed by carina!
carina-jupyterhub
allows you to host a jupyterhub instance where the users bring
their own computing power. You host the jupyterhub frontend, and your
users bring the computing power. Amazing!
To setup your own jupyterhub frontend hosted on carina, create a new cluster on carina and complete the following steps.
- Install and configure the Carina CLI with your Carina username and API key.
- Run
eval $(carina env name-of-your-cluster)
to set up your environment.
- Click "Get Access" on the Carina web interface to download a zip of your credentials.
- Unzip the credentials zip file.
- Run
source docker.env
from the unzipped access file.
docker run --net=host -ti --rm -p 8000 betatim/carina-jupyterhub:25112015 bash
- In the container modify
jupyterhub_config.py
withsed -i -e "s/public_ips()\[0\]/\'`ip addr list eth0 |grep "inet "|cut -d' ' -f6|cut -d/ -f1`'/g" jupyterhub_config.py
This replacespublic_ips()[0]
with the public IP of your cluster injupyterhub_config.py
- start jupyterhub:
jupyterhub
. This will eventually printJupyterHub is now running at http://SomeIPAddress:8000/
. Congratulations, you now runningjupyterhub
on a carina cluster!
To use the service you just built:
- Create a second carina cluster, and download the access credentials using the "Get Access" button in the web UI.
- Visit
http://SomeIPAddress:8000
enter a short username, and upload the credentials file of your second cluster as the Credentials file - Share
http://SomeIPAddress:8000
with your friends (they have to repeat these last three steps)!
While not an immediate threat to your health, please think carefully about the security implications of running untrusted code, giving away credentials to computing resources, etc before doing any of the above.