How To Configure a Continuous Integration Testing Environment with Docker and Docker Compose on Ubuntu 14.04

Updated on November 3, 2016
Default avatar

By Pablo Chico de Guzman

CTO @Okteto Inc

How To Configure a Continuous Integration Testing Environment with Docker and Docker Compose on Ubuntu 14.04

An Article from Docker


Continuous integration (CI) refers to the practice where developers integrate code as often as possible and every commit is tested before and after being merged into a shared repository by an automated build.

CI speeds up your development process and minimizes the risk of critical issues in production, but it is not trivial to set up; automated builds run in a different environment where the installation of runtime dependencies and the configuration of external services might be different than in your local and dev environments.

Docker is a containerization platform which aims to simplify the problems of environment standardization so the deployment of applications can also be standardized (find out more about Docker). For developers, Docker allows you to simulate production environments on local machines by running application components in local containers. These containers are easily automatable using Docker Compose, independently of the application and the underlying OS.

This tutorial uses Docker Compose to demonstrate the automation of CI workflows.

We will create a Dockerized “Hello world” type Python application and a Bash test script. The Python application will require two containers to run: one for the app itself, and a Redis container for storage that’s required as a dependency for the app.

Then, the test script will be Dockerized in its own container and the whole testing environment moved to a docker-compose.test.yml file so we can make sure we are running every test execution on a fresh and uniform application environment.

This approach shows how you can build an identical, fresh testing environment for your application, including its dependencies, every time you test it.

Thus, we automate the CI workflows independently of the application under test and the underlying infrastructure.


Before you begin, you will need:

  • An Ubuntu 14.04 server

  • A non-root user with sudo privileges. You can follow this tutorial to set this up

  • Some familiarity with Docker and Docker Compose; the links are for reference, as we will install the software during this tutorial

  • Testing, testing and more (automated) testing! You should already be convinced about the advantages of testing and applying CI to your development workflows

Step 1 — Install Docker

If Docker is not yet available on your server, the easiest way is to download and execute the official Docker installation script, which prompts for a sudo password (find more information about installing Docker here):

  1. wget -qO- https://get.docker.com/ | sh

In order to make it easier to work with Docker, add your user to the docker group with the following command:

  1. sudo usermod -aG docker $(whoami)

Log out and then log in to your server to activate the docker group for your user.

Step 2 — Install Docker Compose

Docker Compose is an open-source tool for defining and running multi-container applications using a declarative approach. In order to install Docker Compose, execute the following commands:

  1. sudo apt-get update
  2. sudo apt-get -y install python-pip
  3. sudo pip install docker-compose

Verify that docker-compose is correctly installed by executing:

  1. docker-compose --version

You should see something like:

docker-compose version 1.6.2, build 4d72027

This should tell you the docker-compose version you have installed.

Step 3 — Create the “Hello World” Python Application

In this step we will create a simple Python application as an example of the type of application you can test with this setup.

Create a fresh folder for our application by executing:

  1. cd ~
  2. mkdir hello_world
  3. cd hello_world

Edit a new file app.py with nano:

  1. nano app.py

Add the following content:

from flask import Flask
from redis import Redis

app = Flask(__name__)
redis = Redis(host="redis")

def hello():
    visits = redis.incr('counter')
    html = "<h3>Hello World!</h3>" \
           "<b>Visits:</b> {visits}" \
    return html.format(visits=visits)

if __name__ == "__main__":
    app.run(host="", port=80)

app.py is a web application based on Flask that connects to a Redis data service. The line visits = redis.incr('counter') increases the number of visits and persists this value in Redis. Finally, a Hello World message with the number of visits is returned in HTML.

Our application has two dependencies, Flask and Redis, which you can see in the first two lines. These dependencies must be defined before we can execute the application. Edit a new file:

  1. nano requirements.txt

Add the contents:


Step 4 — Dockerize the “Hello World” Application

Docker uses a file called Dockerfile to indicate the required steps to build a Docker image for a given application. Edit a new file:

  1. nano Dockerfile

Add the following contents:

FROM python:2.7


ADD requirements.txt /app/requirements.txt
RUN pip install -r requirements.txt

ADD app.py /app/app.py


CMD ["python", "app.py"]

Let’s analyze the meaning of each line:

  • FROM python:2.7: indicates that our “Hello World” application image is built from the official python:2.7 Docker image
  • WORKDIR /app: sets the working directory inside of the Docker image to /app
  • ADD requirements.txt /app/requirements.txt: adds the file requirements.txt to our Docker image
  • RUN pip install -r requirements.txt: installs the application pip dependencies
  • ADD app.py /app/app.py: adds our application source code to the Docker image
  • EXPOSE 80: indicates that our application can be reached at port 80 (the standard public web port)
  • CMD ["python", "app.py"]: the command that starts our application

This Dockerfile file has all the information needed to build the main component of our “Hello World” application.

The Dependency

Now we get to the more sophisticated part of the example. Our application requires Redis as an external service. This is the type of dependency that could be difficult to set up in an identical way every time in a traditional Linux environment, but with Docker Compose we can set it up in a repeatable way every time.

Let’s create a docker-compose.yml file to start using Docker Compose.

Edit a new file:

  1. nano docker-compose.yml

Add the following contents:

  build: .
  dockerfile: Dockerfile
    - redis
    - "80:80"
  image: redis

This Docker Compose file indicates how to spin up the “Hello World” application locally in two Docker containers.

It defines two containers, web and redis.

  • web uses the current folder for the build context, and builds our Python application from the Dockerfile file we just created. This is a local Docker image we made just for our Python application. It defines a link to the redis container in order to have access to the redis container IP. It also makes port 80 publicly accessible from the Internet using your Ubuntu server’s public IP

  • redis is executed from a standard public Docker image, named redis

Step 5 — Deploy the “Hello World” Application

In this step, we’ll deploy the application, and by the end it will be accessible over the Internet. For the purposes of your deployment workflow, you could consider this to be either a dev, staging, or production environment, since you could deploy the application the same way numerous times.

The docker-compose.yml and Dockerfile files allow you to automate the deployment of local environments by executing:

  1. docker-compose -f ~/hello_world/docker-compose.yml build
  2. docker-compose -f ~/hello_world/docker-compose.yml up -d

The first line builds our local application image from the Dockerfile file. The second line runs the web and redis containers in daemon mode (-d), as specified in the docker-compose.yml file.

Check that the application containers have been created by executing:

  1. docker ps

This should show two running containers, named helloworld_web_1 and helloworld_redis_1.

Let’s check to see that the application is up. We can get the IP of the helloworld_web_1 container by executing:

  1. WEB_APP_IP=$(docker inspect --format '{{ .NetworkSettings.IPAddress }}' helloworld_web_1)
  2. echo $WEB_APP_IP

Check that the web application is returning the proper message:

  1. curl http://${WEB_APP_IP}:80

This should return something like:

<h3>Hello World!</h3><b>Visits:</b> 1<br/>

The number of visits is incremented every time you hit this endpoint. You can also access the “Hello World” application from your browser by visiting the public IP address of your Ubuntu server.

How to Customize for Your Own Application

The key to setting up your own application is to put your app in its own Docker container, and to run each dependency from its own container. Then, you can define the relationships between the containers with Docker Compose, as demonstrated in the example. Docker Compose is covered in greater detail in this Docker Compose article.

For another example of how to get an application running across several containers, read this article about running WordPress and phpMyAdmin with Docker Compose.

Step 6 — Create the Test Script

Now we’ll create a test script for our Python application. This will be a simple script that checks the application’s HTTP output. The script is an example of the type of test that you might want to run as part of your continuous integration deployment process.

Edit a new file:

  1. nano test.sh

Add the following contents:

sleep 5
if curl web | grep -q '<b>Visits:</b> '; then
  echo "Tests passed!"
  exit 0
  echo "Tests failed!"
  exit 1

test.sh tests for basic web connectivity of our “Hello World” application. It uses cURL to retrieve the number of visits and reports on whether the test was passed or not.

Step 7 — Create the Testing Environment

In order to test our application, we need to deploy a testing environment. And, we want to make sure it’s identical to the live application environment we created in Step 5.

First, we need to Dockerize our testing script by creating a new Dockerfile file. Edit a new file:

  1. nano Dockerfile.test

Add the following contents:

FROM ubuntu:trusty

RUN apt-get update && apt-get install -yq curl && apt-get clean


ADD test.sh /app/test.sh

CMD ["bash", "test.sh"]

Dockerfile.test extends the official ubuntu:trusty image to install the curl dependency, adds tests.sh to the image filesystem, and indicates the CMD command that executes the test script with Bash.

Once our tests are Dockerized, they can be executed in a replicable and agnostic way.

The next step is to link our testing container to our “Hello World” application. Here is where Docker Compose comes to the rescue again. Edit a new file:

  1. nano docker-compose.test.yml

Add the following contents:

  build: .
  dockerfile: Dockerfile.test
    - web
  build: .
  dockerfile: Dockerfile
    - redis
  image: redis

The second half of the Docker Compose file deploys the main web application and its redis dependency in the same way as the previous docker-compose.yml file. This is the part of the file that specifies the web and redis containers. The only difference is that the web container no longer exposes port 80, so the application won’t be available over the public Internet during the tests. So, you can see that we’re building the application and its dependencies exactly the same way as they are in the live deployment.

The docker-compose.test.yml file also defines a sut container (named for system under tests) that is responsible for executing our integration tests. The sut container specifies the current directory as our build directory and specifies our Dockerfile.test file. It links to the web container so the application container’s IP address is accessible to our test.sh script.

How to Customize for Your Own Application

Note that docker-compose.test.yml might include dozens of external services and multiple test containers. Docker will be able to run all these dependencies on a single host because every container shares the underlying OS.

If you have more tests to run on your application, you can create additional Dockerfiles for them, similar to the Dockerfile.test file shown above.

Then, you can add additional containers below the sut container in the docker-compose.test.yml file, referencing the additional Dockerfiles.

Step 8 — Test the “Hello World” Application

Finally, extending the Docker ideas from local environments to testing environments, we have an automated way of testing our application using Docker by executing:

  1. docker-compose -f ~/hello_world/docker-compose.test.yml -p ci build

This command builds the local images needed by docker-compose.test.yml. Note that we are using -f to point to docker-compose.test.yml and -p to indicate a specific project name.

Now spin up your fresh testing environment by executing:

  1. docker-compose -f ~/hello_world/docker-compose.test.yml -p ci up -d

Check the output of the sut container by executing:

  1. docker logs -f ci_sut_1
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100    42  100    42    0     0   3902      0 --:--:-- --:--:-- --:--:--  4200
Tests passed!

And finally, check the exit code of the sut container to verify if your tests have passed:

  1. docker wait ci_sut_1

After the execution of this command, the value of $? will be 0 if the tests passed. Otherwise, our application tests failed.

Note that other CI tools can clone our code repository and execute these few commands to verify if tests are passing with the latest bits of your application without worrying about runtime dependencies or external service configurations.

That’s it! We’ve successfully run our test in a freshly built environment identical to our production environment.


Thanks to Docker and Docker Compose, we have been able to automate how to build an application (Dockerfile), how to deploy a local environment (docker-compose.yml), how to build a testing image (Dockerfile.test), and how to execute (integration) tests (docker-compose.test.yml) for any application.

In particular, the advantages of using the docker-compose.test.yml file for testing are that the testing process is:

  • Automatable: the way a tool executes the docker-compose.test.yml is independent of the application under test
  • Light-weight: hundreds of external services can be deployed on a single host, simulating complex (integration) test environments
  • Agnostic: avoid CI provider lock-in, and your tests can run in any infrastructure and on any OS which supports Docker
  • Immutable: tests passing on your local machine will pass in your CI tool

This tutorial shows an example of how to test a simple “Hello World” application.

Now it’s time to use your own application files, Dockerize your own application test scripts, and create your own docker-compose.test.yml to test your application in a fresh and immutable environment.

Thanks for learning with the DigitalOcean Community. Check out our offerings for compute, storage, networking, and managed databases.

Learn more about us

About the authors
Default avatar

CTO @Okteto Inc

Docker Community Leader. Cloud Native Development Advocate.

Still looking for an answer?

Ask a questionSearch for more help

Was this helpful?

This textbox defaults to using Markdown to format your answer.

You can type !ref in this text area to quickly search our full set of tutorials, documentation & marketplace offerings and insert the link!

Hi, I got this error , can you plz help me ?

λ docker logs -f ci_sut_1 sleep: invalid time interval ‘5\r’ Try ‘sleep --help’ for more information. test.sh: line 8: syntax error near unexpected token fi' test.sh: line 8: fi’


Thanks for publishing instruction to setup a to configure a continuous integration testing environment with docker. In addition to the Docker and Docker Compose command line, readers might find useful our GUI for Docker on Ubuntu. It is named Data Pulley. http://datapulley.com Data Pulley is a lightweight management UI which allows you to easily manage Docker containers on your local machine. And, we have just added option to run using Docker Compose as well (File -> Open Docker Compose) Enjoy :)

Thanks for sharing! How do you instruct docker-compose to automatically tear down the redis container after sut tests pass and exit with code 0? Is that job just handled by your ci tool?

should we manually delete the containers once testing is completed ?

Hi! Thank you for the article :) Is it a good idea to extend docker-compose.yml file rather than create a separate docker-compose.text.yml? I have about 8 services in it and files are almost identical. The only difference is in expose params and sut service.

Thank you for this tutorial. I would like to point out something.

WEB_APP_IP=$(docker inspect --format '{{ .NetworkSettings.IPAddress }}' helloworld_web_1)

The above command does not work with the latest version of docker.

Please update it to

WEB_APP_IP=$(docker inspect --format='{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' helloworld_web_1)

Source: docker inspect

Good stuff! Gracias!

Need help:

$ docker-compose -f docker-compose.yml build redis uses an image, skipping Building web Step 1 : FROM python:2.7 —> a3b29970a425 Step 2 : WORKDIR /app —> Using cache —> cd7588fc0a0b Step 3 : ADD requirements.txt /app/requirements.txt —> Using cache —> a769ec7f949d Step 4 : RUN pip install -r requirements.txt ERROR: Service ‘web’ failed to build: open /var/lib/docker/overlay/c68b58dad22103437d82674061a814d165416321cc3856a0f6ce90097737505b-init/merged/.dockerenv: operation not permitted

Thanks in advance!


Surprisingly simple and easy to understand, I had two comments

  1. On my mac I had to use docker-machine to get my IP address (Step 5):

             WEB_APP_IP=`docker-machine ip`
  2. I also ran into the syntax error in the .yml file. Would it be possible to edit the article to remove the erroneous line?

Try DigitalOcean for free

Click below to sign up and get $200 of credit to try our products over 60 days!

Sign up

Join the Tech Talk
Success! Thank you! Please check your email for further details.

Please complete your information!

Get our biweekly newsletter

Sign up for Infrastructure as a Newsletter.

Hollie's Hub for Good

Working on improving health and education, reducing inequality, and spurring economic growth? We'd like to help.

Become a contributor

Get paid to write technical tutorials and select a tech-focused charity to receive a matching donation.

Welcome to the developer cloud

DigitalOcean makes it simple to launch in the cloud and scale up as you grow — whether you're running one virtual machine or ten thousand.

Learn more
DigitalOcean Cloud Control Panel