.. _jupyter:

################################################################################
Accessing Jupyter Notebook launched from a compute node
################################################################################



1. From a terminal window, log in to a CRC front end node (e.g. crcfe01):

.. code-block:: shell

     ssh -Y netid@crcfe01.crc.nd.edu

2. Access a compute node:

For GPU:

.. code-block:: shell

    qrsh -q gpu -l gpu_card=1 -pe smp 1
            
   
For CPU:

.. code-block:: shell

    qrsh -q long -pe smp 1
          
3. Inside the compute node, launch Jupyter notebook:

.. code-block:: shell

    jupyter lab --no-browser --ip='0.0.0.0'

You'd see something like the following :

.. code-block:: shell

    [I 13:52:19.611 NotebookApp] Jupyter Server 2.7.3 is running at:

    [I 13:52:19.611 NotebookApp] http://qa-1080ti-001.crc.nd.edu:8888/?token=4e6fc6f60565415f4220b897f6adad2f3b8b923bb9f77b9c

    [I 13:52:19.611 NotebookApp]  or http://127.0.0.1:8888/?token=4e6fc6f60565415f4220b897f6adad2f3b8b923bb9f77b9c

    [I 13:52:19.611 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).


   
(Please note the server name and its port number. In this example: qa-1080ti-001.crc.nd.edu:8888)

4. Access the Jupyter notebook using SSH tunneling. On your local machine and in a separate terminal window run the following command:

.. code-block:: shell

   ssh netid@crcfe01.crc.nd.edu -L 8888:qa-1080ti-001.crc.nd.edu:8888 -N

(Enter your password. If there are no errors, the command line will hang. This is normal).

5. Open a browser from your local machine and navigate to http://localhost:8888/, the Jupyter Notebook web interface. When prompted, enter token (in this example, token would be 4e6fc6f60565415f4220b897f6adad2f3b8b923bb9f77b9c)