Dockerizing your Applications : Building Images and Working with Registries

In the previous session, we have learnt about various container operations such as running containers from pre built images, port mapping, inspecting and updating containers, limiting resources etc., In this chapter, we are going to learn about how to build containers for your individual applications, as well as how to work with docker hub registry to host and distribute the images.

Lab: Registering with the DockerHub

Since we are going to start working with the registry, build and push images to it later, its essential to have our own account on the registry. For the purpose of this tutorial, we are going to use the hosted registry i.e. Dockerhub.

Steps to create Dockerhub account

Step 1:

Visit the following link and sign up with your email id


Step 2:

Check your email inbox and check the activation email sent by docker team

Step 3:

After clicking on the activation link, you will be redirected to a log in page. Enter your credentials and log in


You will be launched to Dockerhub main page. Now the registration process is complete and you have account in Dockerhub!


Lab: Building Docker Images - A manual approach

Before we start building automated images, we are going to create a docker image by hand. We have already used the pre built image from the registry in the last session. In this session, however, we are going to create our own image with ghost installed. Since Ghost is a node.js based application, we will base our work on existing official image for node

Clone Repository for Java worker app

git clone

Launch a intermediate container to install worker app

Create a Container with schoolofdevops/voteapp-mvn:v1 image

docker run -idt --name interim schoolofdevops/voteapp-mvn  sh

Copy over the Source Code

cd example-voting-app/worker
docker container cp .  interim:/code

Connect to container to compile and package the code

docker exec -it interim sh

mvn package

Verify jarfile has been built

ls target/

java -jar target/worker-jar-with-dependencies.jar

[sample output]

/code # java -jar target/worker-jar-with-dependencies.jar
Waiting for redis
Waiting for redis
Waiting for redis
Waiting for redis
Waiting for redis
Waiting for redis

[use ^c to exit]

The above is the expected output. The worker app keeps waiting for redis and then later db in a loop.

Move the artifact, remove source code

mv target/worker-jar-with-dependencies.jar /run/worker.jar

rm -rf /code/*


Commit container to an image

  • Exit from the container shell
  • Note container ID

Commit the container into a image as,

docker container commit interim  <docker hub user id >/worker:v1

Test before pushing by launching container with the packaged app

  docker run --rm -it  <docker hub user id >/worker:v1 java -jar /run/worker.jar
Push Image to registry

Before you push the image, you need to be logged in to the registry, with the docker hub id created earlier. Login using the following command,

docker login

To push the image, first list it,

docker image ls

[Sample Output]

REPOSITORY                   TAG                 IMAGE ID            CREATED             SIZE
initcron/worker         v2              90cbeb6539df        18 minutes ago      194MB
initcron/worker         v1              c0199f782489        34 minutes ago      189MB

To push the image,

docker push <dockrhub user id>/worker:v1

Lab: Building Images with Dockerfile

Now, lets build the same image, this time with Dockerfile. To do this, create a file by name Dockerfile in the root of the source code.

file: example-voting-app/worker/Dockerfile

FROM schoolofdevops/maven


COPY .  .

RUN mvn package && \
    mv target/worker-jar-with-dependencies.jar /run/worker.jar && \
    rm -rf /app/*

CMD java  -jar /run/worker.jar

Lets now build the image

cd example-voting-app/worker

docker image build -t <dockrhub user id>/worker:v2 .

docker image ls

Try building again,

docker image build -t <dockrhub user id>/worker:v2 .

This time, it does not build everything, but uses cache.

Testing the image

docker container run --rm -it  <dockrhub user id>/worker:v2

Tag the image as latest,

docker image tag  <dockrhub user id>/worker:v2  <dockrhub user id>/worker:latest

docker image ls

Finally, publish it to the registry,

docker image push <dockrhub user id>/worker:latest

docker image push <dockrhub user id>/worker


Building Base Images: