Jupyter Notebook at Hemera

You can request and access your Jupyter notebooks on hemera at:

In order to use Jupyter Notebook sessions You need to have access to the HPC systems. You can request it by sending an e-mail to cluster-admin@hzdr.de.

When using the links below, please reload the following page until the link with the job id turns active, then click on it.

The dashboard displays a link below the submit button which always points to the job settings you currently entered (requires JavaScript). You can bookmark this link to save you favorite settings, like with two examples above.

Unsaved data will be lost when the job ends and the Jupyter server is shut down. Save your work. Use ctrl+s.

The dashboard also provides Pluto.jl (Julia) notebooks, which can be selected from the drop-down list defaulting to "Jupyter".

Usage of python virtual environments in notebooks

To use a pyton virtual environment in a notebook, the notebook's kernel needs to be set to a kernel inside that environment, independently of whether a custom module environment is loaded using a jupyter.profile.

After creating and loading the desired envionment, perform the follwing steps inside the venv:
pip install ipykernel
python -m ipykernel install --user --name=myenv
Now, the new kernel inside myenv should be available to select in jupyter.

Access from outside the HZDR Network

This service can only be accessed directly from within the HZDR network, which should include VPN. However, you can reach it from the outside by using uts.hzdr.de as a proxy which you need to configure in your browser. For this you need to set up a SOCKS proxy via SSH:

  • on a Linux client: on the commandline, run e.g. ssh -D 2000 uts.hzdr.de or ssh -D 2000 mup.hzdr.de
  • on a Windows client PuTTY provides this functionality, one may follow this tutorial.
You can then use this proxy by configuring your browser to access the internet via this local proxy server. (e.g.: Host=localhost Port=2000, Firefox, Chrome )

Loading your own Environment

You can cause the service to load your own environment by placing a file named jupyter.profile in your home directory or your intended working directory. Before Juypter is started, any such file in the working directory selected in the dashboard will be sourced, if there is none, any jupyter.profile in your home directory is be used. Otherwise the default environment is loaded. In any case a Python 3 module which provides a jupyter installation will be loaded first. If you choose to unload or replace the pre-loaded module python/3.10.4-env_a make sure, that a jupyter notebook server is still available in your environment. Note: If you are activating a Python virtual env in your jupyter.profile, that virtual env should contain its own jupyter. If the system-juypter is used, it will run outside of your venv, even though started inside it.