Docker demo

Table of Contents

• How to run the project
  • Create the virtual environment
  • Activate the virtual environment
  • Deactivate the virtual environment
• How to build a Docker image: docker build -t name_of_image:version .
• How to create a Docker container from a docker image
• Demo of Common errors with creating Docker files
  • Always specify the base image version tag instead of latest.
  • Order of Docker instructions matters
  • Impact of a .dockerignore file

How to run the project

1. Create the virtual environment

python -m venv venv

1. Activate the virtual environment

   • for windows

   venv\Scripts\activate

   • for Linux / macOS

   source venv/bin/activate

2. Deactivate the virtual environment

   deactivate

• Install project dependencies

• Run the project by running

  python -m app.py

How to build a Docker image: docker build -t name_of_image:version -f name_of_dockerfile .

docker build -t hello_world:1.0 .

You can leave out the -f flag if your docker file is so named Dockerfile and is in the current

or

docker build -t hello_world:1.0 -f Dockerfile  .

How to create a Docker container from a docker image

docker run --name hello-udacians -p 5500:5000 -e FLASK_ENV=development -e FLASK_DEBUG=True hello_world:1.0

• -p flag publishes the application running on port 5000 within the docker container to be accessed via port 5500 on the host machine
• -e flag is used to specify environment variables
• --name is used to specify the name of the docker container, without it a random name is assigned to the container.

Demo of Common errors with creating Docker files

Always specify the base image version tag instead of latest.

The file broken.Dockerfile builds on the latest python version using broken_requirements.txt file. The broken_requirements.txt file is called so for lack of a better name but it contains an application's requirements that runs on python 3.6 Try building the docker image by running

docker build -t image_build_will_fail -f broken.Dockerfile .

As you can see, the build does not pass. When you change the image tag from latest to 3.6 to point to a specify python version 3.6, the build will pass

docker build -t image_build_will_succeed -f broken.Dockerfile .

Lesson learnt: Always specify image tag version to a specific one to make your project fail proof in the future with regards to the environment in which it successfully runs.

Order of Docker instructions matters

The wrong_order.Dockerfile has the instruction for installing requirements appearing before the COPY instruction that copies the reqiuirements.txt file.

docker build -t build_failed_wrong_order -f wrong_order.Dockerfile .

Lesson learnt: Always ensure that you declare the docker instructions in the right order.

Impact of a .dockerignore file

Build the project again

docker build -t HelloWorld:1.0  -f complete.Dockerfile .

Take note of the size of build context and the final image Delete the .dockerignore file and build the project again

docker build -t HelloWorld:1.1  -f complete.Dockerfile .

Take note of the size of the build context and the final image Run docker images Compare size of image HelloWorld:1.0 with HelloWorld:1.1

HelloWorld:1.1 had a bigger build context and also larger image size than HelloWorld:1.0

Lesson learnt: The .dockerignore file helps us exclude unwanted files from the build context and as a result we are going to have a smaller build context (the set of files located in the specified PATH or URL) reusulting and smaller image size as well since the COPY instruction will have a much smaller scoped reference to a file in the context.

This makes a difference if you are copying all files like it is with the complete.Dockerfile whereas with the simple Dockerfile provided in the repo where you specify individual files to copy, only the build context will be affected but the image size will not change since there's no avenue for copying any unnecessary files. Bigger projects usually have many files and copying them one by one is against the convention and would result in more layers. Remember every instruction in a dockerfile is a layer.