Skip to content

Standalone jupyter notebooks

Access to Maxwell or any other advance compute resource is actually not needed at all to run a jupyter notebook. It's possible to launch a private notebookserver acting pretty much like a jhub session, but you will need access to the port your private notebook server is listening to. The simplest options are

  • use a FastX session on naf or maxwell login nodes
  • start the notebook and the browser directly on the login node. If running out of memory
    • allocate an interactive batch node
    • ssh into the batch node and start your notebook server on the batch node
    • the browser in the FastX session will have access to the ports needed

Configuring and launching a notebook server

for simplicity there is a primitive script configuring and launching a standalone notebook server:

/software/tools/bin/jupyter-standalone.sh

The script will offer two locations like

   http://max-wgse001.desy.de:3904/tree?token=83389f
   http://127.0.0.1:3904/tree?token=83389f

Use the first line with full hostname (copy&paste in your browser, or CTRL+left-mouse to open it). Rendering might take a moment, so might show a grey screen for a couple of seconds.

Have a look at the script to customize according to your needs.

Note: the server can not be accessed from non-DESY networks.