JupyterHub is a multi-user server that allows users to access and run Jupyter Notebooks through a web browser. https://jupyterhub.apps.stack.it.ntnu.no/ is a Jupyterhub JupyterHub installation that is open and available to for use for everyone at NTNU. The hub can be used for teaching or academic personal use. JupyterHub at NTNU is maintained by NTNU IT Forskningsstøtte.
In addition we host a number of custom dedicated hubs for different courses at NTNU, usually found under {course code}.apps.stack.it.ntnu.no. Access to these dedicated hubs is generally limited to students and instructors in the specified courses. However, almost everything mentioned in this wiki also applies to these dedicated hubs. Instructors can find information about getting access to a hub for their course in the article JupyterHub for teachingThe environment has a limited amount of resources, so it is not necessarily suitable for heavier workloads. The hub does not have access to GPUs. For heavier computation needs, please look into using for example IDUN.
Basic usage
You can access JupyterHub at https://jupyterhub.apps.stack.it.ntnu.no/. Log in with your normal NTNU account through Feide.
On the hub you are limited to a maximum of 1 CPU core and 1 GB of memory for running your notebooks. You have access to persistent storage
, so- everything under your home directory ~/ (same as /home/jovyan/) - meaning that all your files stored under your home directory will be stored between sessions. NOTE: Admin users have full access to each users' environment, so please do not store sensitive data on the hub!
The hub has a limited amount of resources, so it is not necessarily suitable for heavier workloads. The hub does not have access to GPUs. For heavier computation needs, please look into using for example IDUN.
The JupyterHub user environment is based on the datascience-notebook image from Jupyter , and supports the Python, R and Julia programming languages by default. Specifically for Python, it the hub comes with many of the commonly used datascience data science and visualisation packages pre-installed, such as numpy, pandas, scipy, sklearn, matplotlib, seaborn... just to mention a few. You can also freely install any packages you might need yourself.
Each users' Jupyter server and kernels are run in a full-featured linux-based container. The container part means that each users' enviroments are completely isolated from everyone else - even if a single user manages to crash or break their own environment, other users will not notice any difference. The linux part means that you have access to most common UNIX commands as well as many useful non-Jupyter related tools, such as git, through the terminal.
Your Jupyter server is automatically stopped after 20 minutes of inactivity, or after 24 hours of maximum run time. You can always simply restart your server through the hub if you run into these limits.
Jupyterlab / Notebook
You can always choose whether you want to use Jupyterlab or Jupyter Notebook interfaces. The hub defaults to Jupyterlab, but you can change to notebook simply by changing the end of the URL in your browser after logging in.
Jupyterlab: https://jupyterhub.apps.stack.it.ntnu.no/user/{your_username}/lab
Notebook: https://jupyterhub.apps.stack.it.ntnu.no/user/{your_username}/tree
Contact us
If you have any questions or problems, feel free to contact us through NTNU Hjelp or send us an email at jupyterhub (at) it.ntnu.no.
Students' first line support should always first be the course staff - they will relay relevant questions to us.