Workflow

This is a workflow to show how to open jupyter notebook on CRI or RCC. There are some detailed instructions here, CRI’s instructions and More general instructions

Remote side

Make sure you’ve installed the required Python modules and packages.

$ ssh <username>@gardner.cri.uchicago.edu

$ jupyter notebook --no-browser --port=XXXX

The default port will be 8888, and “–no-browser” is required because if you do not specify –no-browser –ip=, the web browser will be launched on the node and the URL returned cannot be used on your local machine.

If you set port=8889, the result should be like the following:

To access the notebook, open this file in a browser:
file:///home/<username>/.local/share/jupyter/runtime/nbserver-129406-open.html
Or copy and paste one of these URLs:
http://localhost:8889/?token=47871f1a27e41715e04362540f5730611a30b17ae2072827
or http://127.0.0.1:8889/?token=47871f1a27e41715e04362540f5730611a30b17ae2072827

Local machine

Open a new terminal and run the following command.

$ ssh -N -f -L localhost:YYYY:localhost:XXXX <username>@gardner.cri.uchicago.edu

Select any port=YYYY which you haven’t used for other work.

Then open a browser, type in localhost=<YYYY> and copy and paste the token from the last step.
Finally, it should worked out.