Docker PowerShell Jupyter Notebook Environments

Previously in this post I detailed getting started with Local PowerShell Jupyter Notebooks. In this post I show how to take an existing Docker Community Reference CSharp Jupyter Stack Container Image and update it for Dot Net Interactive and PowerShell to enable containerized Docker PowerShell Jupyter Notebook. I also show how to automate the build of that Docker Image using a Dockerfile, and how to run that PowerShell Jupyter environment in Binder building from the Dockerfile via my GitHub Repo. Finally if you just want a Docker Container Image, I’ve also published it to Docker Hub here.

You may also be interested in my Microsoft Reactor session “Elevate your documentation with PowerShell Jupyter Notebook“.

1. Local Manual Docker PowerShell Jupyter Notebook


  • You will need Docker installed locally and configured for Linux Containers
  • You will need ~4.5Gb of available diskspace.

Process Overview

  • pull the Community Reference CSharp Jupyter Container Image
  • update the image for;
    • PowerShell 7
    • Dot Net Core 3.1 SDK
    • Dot Net Interactive
  • Create a Jupyter Notebook Start Script
  • Save and Commit the Image
  • Start the Image

Pull Community Reference Image

As per the introduction we will start with a Docker Image that is already a community stack reference image for CSharp as it already has Jupyter and its dependencies. The tlinnet/csharp-notebook container image uses Ubuntu 18.04 so all the dependencies we need to install for PowerShell Jupyter Notebook will be for Ubuntu 18.04.

docker pull tlinnet/csharp-notebook

CSharp Jupyter Container Image

Start the Image

With the image now local let’s start it with root privileges so we can update it.

docker run -it -u root --name pwsh-jupyter tlinnet/csharp-notebook bash

Install PowerShell

First up we will want to install PowerShell. This will install the latest version of PowerShell (currently 7.0.2).

wget -q 
sudo dpkg -i packages-microsoft-prod.deb
sudo apt-get update
sudo apt-get install -y powershell

Install PowerShell 7 on Ubuntu 18.04

Install Dot Net Core SDK

Next, we will need to install the latest Dot Net Core SDK.

sudo apt-get install -y apt-transport-https && sudo apt-get update && sudo apt-get install -y dotnet-sdk-3.1

Install Dot Net SDK on Ubuntu 18.04

Dot Net Interactive

Then we need to install Dot Net Interactive.

dotnet tool install -g Microsoft.dotnet-interactive --version 1.0.136005 --add-source ""

Install Dot Net Interactive on Ubuntu 18.04

Configure environment paths for .dotnet tools

The following is required for the future when running the container image. We need to add the path to .dotnet/tools.

Edit the ~/.bashrc file to include the path to .dotnet tools. Use nano to update .bashrc

nano ~/.bashrc
# Add path to .NET Tools: 
export PATH=$PATH:~/.dotnet/tools

Path to DotNet Tools on Ubuntu 18.04

Control + x and Y to save.

Register Dot Net Interactive for Jupyter

We then need to register the Dot Net Interactive Kernels for Jupyter.

export PATH=$PATH:~/.dotnet/tools
dotnet interactive jupyter install

Register dotnet interactive jupyter on Ubuntu 18.04

Validate our Jupyter Kernels

We can then validate our .NET kernels were successfully registered.

jupyter kernelspec list

Validate PowerShell .Net Juptyer Kernels on Ubuntu 18.04

Create a Jupyter Notebook Start Script

The final part of the configuration is to use nano again, to create a small bash script to start Jupyter Notebook.

export PATH="$PATH:/home/jovyan/.dotnet/tools"
jupyter notebook --allow-root

Start Jupyter Notebook with PowerShell on Ubuntu 18.04

Control + x and Y to save.

Stop the Image and update the Container Image

With our customizations complete, we can stop the image and save our Container Image.

docker stop pwsh-jupyter
docker commit pwsh-jupyter pwsh-jupyter

Stop and update the Jupyter PowerShell Docker Container Image

Start PowerShell Jupyter Docker Image

We can then start our Container Image and start Jupyter Notebook. We will also take our localhost home directory and make it available inside the container (so you can load and save your Notebooks to the localhost). Start the image from a command prompt path that is a parent for you home directory or working documents files.
Note: If this is the first time you’ve done this, you will get a prompt from Docker requesting access to your local hard disk.

docker run -it --rm -p 8888:8888 -v "%CD%":/home/jovyan/work pwsh-jupyter bash ./

Start Jupyter PowerShell Notebook Container Image on Ubuntu 18.04

Copy the localhost URL with token key from the console window and paste it into your browser and the Jupyter dashboard will load.

e.g. from the screenshot above the URL is:


Jupyter Notebook Started on Ubuntu 18.04

From the New menu we now have the .NET interactive kernels. We can choose .NET (PowerShell) and start a new PowerShell Jupyter Notebook.

New Jupyter PowerShell Notebook

We can validate our PowerShell environment by running the $psversiontable PowerShell command.

Jupyter PowerShell Environment

2. Build a Container Image from Dockerfile

The above method is all well and good, but what if you need to do it on multiple machines or in the cloud? For that we will need to automate it. Clone my PowerShell Jupyter Notebook Dockerfile Repo from . With the zip file downloaded from the GitHub link above and expanded you will have the Dockerfile and the config directory that contains the startup script.

Download PowerShell Jupyter Notebook Dockerfile

Build the Docker Image from Dockerfile

docker build --tag darrenjrobinson/pwsh-jupyter-notebook .

Build PowerShell Jupyter Notebook Docker Image from Dockerfile

Run the Docker Image

docker run -it --rm -p 8888:8888 -v "%CD%":/home/jovyan/work darrenjrobinson/pwsh-jupyter-notebook:latest bash ./

Run PowerShell Jupyter Notebook Docker Image

You can then navigate to the instance using a browser and the URL shown in the session.

3. Run a Docker Container Image in Binder

Binder is an online service for running Jupyter Notebook. The link below will spin up a Docker Container with the image detailed in this post.

Start Binder Jupyter PowerShell Notebook Container Image on Ubuntu 18.04

Once started you can start a New PowerShell Notebook and start powershelling.

New Binder Jupyter PowerShell Notebook Container Image on Ubuntu 18.04

4. Docker Hub

Here is the link to my published Docker PowerShell Jupyter Image on Docker Hub.


In this post I’ve shown how to take a community reference Jupyter Notebook image and update it to enable PowerShell Jupyter Notebook. I’ve then shown how to automate the generation of the Docker Container Image and finally how to run my PowerShell Jupyter Notebook online using Binder. Enjoy.