Setting Up Jupyter Notebooks

The Jupyter Notebook is an open-source web application that allows you to create and share documents that contain live code, equations, visualizations and narrative text. Uses include: data cleaning and transformation, numerical simulation, statistical modeling, data visualization, machine learning, and much more. The following instructions will explain how to setup a jupyter notebook securely.

1. ssh into the cluster

   •  ssh -X -v <userID>@sporcsubmit.rc.rit.edu
     

2. Start an [sinterative][submitting_a_job_with_sinteractive] job with the configuration that you will need (e.g. memory, cores, time, gres)

3. Load the packages you need.
   • If you have an environment, you need can run spack env activate <environment>.
   • If you don’t have an environment, do your normal spack load <package> for the packages you need.
4. Load jupyter notebook

   •  spack load py-notebook/hfaavup # or the version you require
     

   •  spack find py-notebook # to find other versions
     

5. Start jupyter notebook

   •  $ jupyter notebook --ip=0.0.0.0 --no-browser
     

6. You will see output similar to this:

   •  [demo@skl-a-47 ~]$ jupyter notebook --ip=0.0.0.0 --no-browser
      [I 10:47:16.816 NotebookApp] The port 8888 is already in use, trying another port.
      [I 10:47:16.820 NotebookApp] Serving notebooks from local directory: /home/slpits
      [I 10:47:16.820 NotebookApp] The Jupyter Notebook is running at:
      [I 10:47:16.820 NotebookApp] http://skl-a-47.rc.rit.edu:8889/?token=803f723f3d20e5e93d24af7073c822c3e77387ce7e15e5cd
      [I 10:47:16.820 NotebookApp]  or http://127.0.0.1:8889/?token=803f723f3d20e5e93d24af7073c822c3e77387ce7e15e5cd
      [I 10:47:16.820 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
      [C 10:47:16.823 NotebookApp]
     

   •    To access the notebook, open this file in a browser:
      file:///home/slpits/.local/share/jupyter/runtime/nbserver-27411-open.html
        Or copy and paste one of these URLs:
      http://skl-a-47.rc.rit.edu:8889/?token=803f723f3d20e5e93d24af7073c822c3e77387ce7e15e5cd
        or http://127.0.0.1:8889/?token=803f723f3d20e5e93d24af7073c822c3e77387ce7e15e5cd
      [I 10:56:29.944 NotebookApp] 302 GET /?token=803f723f3d20e5e93d24af7073c822c3e77387ce7e15e5cd (129.21.171.201) 0.50ms
     

7. There are a few things you need from that output:
   • The port number that jupyter is using, which in this case is 8889
   • The node that sinteractive is running on, which in this case is skl-a-47.rc.rit.edu
   • The URL you need for localhost, which in this case is http://127.0.0.1:8889/?token=803f723f3d20e5e93d24af7073c822c3e77387ce7e15e5cd
8. From a new local terminal window, ssh into sporcsubmit using the -L flag:
   • ssh <userID>@sporcsubmit.rc.rit.edu -L <port_number>:<sinteractive_node>:<port_number>
   • Fill in <port_number> and <sinteractive_node> from Step 7.
9. To access the notebook, paste the URL for localhost (from Step 7) in a new web browser tab:
   • http://127.0.0.1:8889/?token=803f723f3d20e5e93d24af7073c822c3e77387ce7e15e5cd