This space serves as a knowledge base for jupyterhub.uni-muenster.de, a platform for interactive data analysis provided by the University of Münster.
If you have further questions regarding JupyterHub, please join us on Mattermost: mattermost.uni-muenster.de/wwu/channels/jupyterhub.
The Jupyterhub is available for every member of the university, including employees and students. The resource quotas depend on the respective user's status and user groups. Higher resource quotas up to 16 vCPUs and 64GB of RAM and additional GPUs can be requested via email at email@example.com.
We offer two different configurations:
- eScience: Mainly for natural science and data analysis
- Development: Mainly for programming, e.g. C++, Rust
Cull inactive sessions
To fairly distribute resources, we will end idle session after 60 minutes of inactivity for students and 240 minutes for employees. Additionally, we might also stop non-idle sessions, if they run for more than 24h, e.g., in case of security relevant updates to the cluster. Jupyterhub is designed for interactive session. If you require longer running non-interactive jobs or more compute power, the HPC cluster will be a better option (more information at their website.)
Installing your own software
Although many important packages are preinstalled into the images, you can install additional packages by yourself.
If you want to add python packages, you have two options:
- Pip: Within a python notebook you can execute
!pip install package_nameto install the package package_name
- Conda: Since most packages are installed in read-only folders, you have to create a new environment (see below)
Additionally, you can download and execute any file with user privileges using the terminal. With the X11 terminal (Konsole), you can also access GUI applications.
Create new conda environment as a kernel in JupyterHub
To install new packages you have to initialize your shell. To do so, open the terminal and execute:
conda init bash
To create a new environment with Python 3.11 and the package pandas under the name py311:
conda create --name py311 python=3.11 ipykernel pandas # ipykernel is required for jupyterhub conda activate py311 python -m ipykernel install --user --name py311 --display-name "Python (py311)"
After a reload of the page, you should be able to select your new eniroment as a kernel.
Your personal home folder
~ is also available under
\\wwu.de\ddfs\Cloud\wwu1\u_jupyterhub\home within the university network. It can be accessed via the file explorer in Windows or mounted in unix systems.
Cloud storage is mounted at
/cloud/wwu1. You can create a symlink to your project folder in your home directory by executing
ln -s /cloud/wwu1/my_cloud_folder ~ . This allows you to use the file browser on the left hand side to access the data.
All employees of the university are welcome to contribute to this knowledge base and can request write access for confluence by contacting Christian Schild (Mattermost: @schild oder E-Mail: firstname.lastname@example.org).