JupyterHub-Docker
This repository contains pre-set Littlest JupyterHub docker image that can be used for instant JupyterHub deloyment at any VM that runs Ubuntu.
Features
- Instant & simple deployment
- Jupyter Notebook import
Test server
Go to http://79.137.202.188/ to see JupyterHub-Docker in action.
Adding a domain was too complicated as it only works with HTTPS.
If I find it out, the server will also be available at docker.ambassador4ik.dev
Preparation
Install Git:
sudo apt-get update && sudo apt-get install git -y
Clone repo:
git clone https://github.com/Ambassador4ik/JupyterHub-Docker.git
Quick Start (Ubuntu only)
Make sure that you are running Ubuntu 18.04 or higher. Quick start shell script only works on ubuntu.
Go to scripts folder:
cd JupyterHub-Docker/scripts
Finally, run quick setup script:
sudo bash quick-start.sh
Manual Install (Any system)
Make sure that you have installed docker. If not, go here and follow suggested steps for your system.
Now that you have docker installed, you need to build the image.
To do this, cd
to /JupyterHub-Docker
folder and run the following command:
docker build -t quick-jpnb-docker .
You can replace quick-jpnb-docker
with any other name you'd like.
To run built image, execute the following command:
docker run -it \
-p 80:8000 \
-v $PWD/notebooks:/home/admin/notebooks \
--name jpnb-docker \
quick-jpnb-docker
To edit JupyterHub port, replace 80
with any other port number you need.
To edit notebooks import directory, replace $PWD/notebooks
with your path.
Also, make sure to replace quick-jpnb-docker
with the name you created whem building image.
Importing notebooks
To import notebooks, you can simply put them into /JupyterHub-Docker/notebooks
folder.
See Manual Install section for info on custom location usage.
Remarks
The first thing I tried was creating an image from pure Ubuntu, which was not a huge success. You can see the results in /legacy
folder. It kind of works, but only in single-user mode (resque-target
on boot), which is not a good thing at all.