Jupyter Workflow

This chapter describes the recommended workflow for using JupyterLab on the server.

Overview

The workflow consists of five steps:

  1. Create a Python virtual environment (only once).

  2. Start a Jupyter server on the remote machine.

  3. Create an SSH tunnel from the local machine.

  4. Open JupyterLab in a local web browser.

  5. Upload and execute notebooks.

Step 0: Create a Dedicated Jupyter Environment

This step is performed only once.

Create a virtual environment:

python3 -m venv ~/venvs/jupyter

Activate it:

source ~/venvs/jupyter/bin/activate

Install JupyterLab:

pip install jupyterlab

Verify the installation:

jupyter lab --version

Step 1: Start the Jupyter Server

Connect to the server:

ssh username@server.example.edu

Activate the environment:

source ~/venvs/jupyter/bin/activate

Launch JupyterLab:

jupyter lab --no-browser --ip=127.0.0.1 --port=8888

A message similar to the following will appear:

http://127.0.0.1:8888/lab?token=<TOKEN>

Keep this terminal open.

The Jupyter server will continue running in this session.

Step 2: Create an SSH Tunnel

Open a second terminal on your local computer.

Create a tunnel:

ssh -L 8888:127.0.0.1:8888 username@server.example.edu

After authentication, leave this terminal open.

The local port 8888 is now connected to the Jupyter server running on the remote machine.

Step 3: Open JupyterLab

Open a web browser and visit:

http://localhost:8888/tree

If prompted, paste the token displayed when the Jupyter server was started.

Alternatively, use the complete URL displayed by Jupyter:

http://127.0.0.1:8888/lab?token=<TOKEN>

You should now see the JupyterLab interface.

Step 4: Upload Files

JupyterLab provides a graphical interface for managing files.

To upload a notebook or script:

  1. Open the file browser panel.

  2. Navigate to the desired directory.

  3. Click the Upload Files button.

  4. Select one or more files from your computer.

  5. Wait until the upload completes.

Alternatively, files can be dragged directly into the file browser.

After the upload finishes, the files become immediately available on the server filesystem.

Opening Notebooks

Double-click a notebook file:

*.ipynb

to open it in JupyterLab.

Cells can be executed individually or sequentially.

Kernel Management

To restart a kernel:

Kernel → Restart Kernel

To interrupt a computation:

Kernel → Interrupt Kernel

Stopping Jupyter

Return to the terminal where Jupyter was launched.

Press:

Ctrl+C

and confirm the shutdown.