2.1 Dockerfiles

The most common way to create an image is to use a Dockerfile, a text file that contains all the instructions necessary to build the image. The advantage of the Dockerfile is that it can be interpreted by the Docker engine, which makes the creation of images an automated and repeatable task.

Suppose that we want to create a containerized application to download figures from a Web page. As a template for this application, we need to build a new image, that we’ll call fig-downloader.

The Dockerfile containing the instructions to build the image fig-downloader is as follows:

FROM ubuntu
RUN apt-get update
RUN apt-get install -y wget
RUN mkdir -p /my-figures
WORKDIR /my-figures
ENTRYPOINT ["wget", "-nd", "-r", "-A", "jpg,jpeg,bmp,png,gif"]
CMD ["https://www.centralesupelec.fr/fr/presentation"]

Here’s the explanation:

1. We use the image ubuntu as the base image. This corresponds to the instruction FROM ubuntu.

2. We install the utility wget in the base image. This corresponds to the instructions RUN apt-get update and RUN apt-get install -y wget.

3. We create a directory my-figures under the root directory of the image. This corresponds to the instruction RUN mkdir -p /my-figures.

4. We set the newly created directory /my-figures as the working directory of the image. This corresponds to the instruction WORKDIR /my-figures.

5. We specify the command to be executed when a container is run from this image. This corresponds to the instruction

ENTRYPOINT ["wget", "-nd", "-r", "-A", "jpg,jpeg,bmp,png,gif"]

This instruction means: execute wget with the options -nd, -r, -A; the last option takes a list of file extensions (jpg,jpeg,bmp,png,gif) as its argument.

6. Remember that the utility wget takes the URL of the Web page as an argument. The URL will be specified when we run the container from the image fig-downloader. Optionally, we can specify a default argument by using the keyword CMD. The meaning of the instruction:

CMD ["https://www.centralesupelec.fr/fr/presentation"]

is: if we don’t give any URL when we run the container, the figures will be downloaded from https://www.centralesupelec.fr/fr/presentation.

2.2 Building an image

We’re now going to build an image from a Dockerfile.

1. Create a directory named fig-downloader in your computer with a file named Dockerfile inside.

2. In the Dockerfile write the set of instructions that you proposed in Exercise 2.2.

3. In the terminal, set the working directory to fig-downloader.

4. Build an image called fig-downloader by executing the following command:

docker build -t fig-downloader .

The . at the end of the command means that the Docker engine will look for a file named Dockerfile in the working directory.

In order to verify that the new image has been created, type the following command:

docker images

docker run --name dl-2 fig-downloader https://www.centralesupelec.fr/

2.3 Containerized Python application

Download this archive file and unzip it into your working directory. In this archive you’ll find:

• A Dockerfile.
• A Python script main.py that asks the user to enter the URL and the language of a Web page, and prints the 10 most frequent words occurring in that page.
• A file requirements.txt with the list of the Python packages needed to run the given script.

The content of the Dockerfile is as follows:

FROM python:3.7-slim
RUN mkdir -p /app
WORKDIR /app
COPY ./main.py ./requirements.txt /app/
RUN pip install -r requirements.txt
ENTRYPOINT ["python", "main.py"]

Play with the application by running the following command:

docker run --rm -it wordfreq

The containerized application will prompt you to insert the URL of a webpage and the language of the page (in English). The output will be the 20 most used words in the webpage.

3.1 Docker volumes

A volume can be seen as a virtual storage device attached to a container. All files there are written to a volume survive the containerized application that created them. In other words, when a container is destroyed, all the files created by the application in the container remain in the volume.

Let’s create a new Docker volume called data-volume:

docker volume create data-volume

4.1 Creating networks

By default, any new container is attached to the network named bridge.

In order to create a new network, you can use the following command:

docker network create network_name

4.2 Communication between containers

Let’s see if and when the three containers can communicate.

You can now exit the three containers.