Docker Jupyter Lab

Dockerfile and docker-compose files to build a Jupyter Lab image and run docker containers.

Build from dockerfile

To build the image from the Dockerfile, simply run the following line in the terminal:

$ docker build -t jupyter-lab --build-arg JUPYTER_PASS=StrongPassWord .

• -t is the option to add labels to the built image. For example jupyter-lab.
• --build-arg is the option to add values to arguments in build time. In this case, it is required to include a value to JUPYTER_PASS which will be the password to enter the jupyter workspace.

Subsequently, it is possible to change the password. This is done by opening a terminal within the jupyter working environment and executing the following:

$ jupyter lab password

Run

To run a container with the previously constructed image, simply execute the following command:

$ docker run -it -d --name jupyter -p 9000:9000 jupyter-lab

• -it allows the container to start in interactive mode.
• -d runs the container without attaching to it.
• --name gives the container a custom name.
• -p allows you to connect port of the localhost to port of the container. By default, the jupyter lab server listens on port 9000.

Docker Compose

For lauch a service of Jupyter Lab, execute the following command:

$ docker compose up -d

the above command will build the image from the Dockerfile and run a container.

The default jupyter lab login password is abc123, which can be changed directly in the docker-compose.yml file or in the console inside the jupyter lab workspace.

To stop the service, run:

$ docker compose stop

To start the service, run:

$ docker compose start

Finally, to stop and remove the service, execute:

$ docker compose down

Remote connection via SSH

If you run the containers on a remote server and need to work locally in the jupyter lab workspace, you need to access via SSH tunneling.

This is achieved by running on a terminal in the computer where you are going to work:

$ ssh -fN user@server -L 9000:localhost:9000

• -f sends the request to be executed in the background.
• -N runs remote connection without attaching to remote console.
• -L bind the local computer port to the remote server port where the jupyter lab service is listening.

Finally, we launch a web browser, access localhost:9000 and enter the Jupyter Lab login page.

We're ready to work!