About the service
The service was a heat pump performance evaluation tool writen in Python as Web Application using Bokeh framework. The client decided to use ECS because of the simplicity of the deployment and the ability to use Fargate as a compute engine. They packed the application in a Docker container and pushed it to ECR manually. My task was to automate the deployment process using CodePipeline. I started by creating a new CodePipeline using the AWS Console. I chose CodeCommit as a source provider and ECS as a deployment provider, the ECS cluster was already created by the client.The build part
The build stage was pretty simple, I used CodeBuild to build the Docker image and push it to ECR. The buildspec.yml file looked like this:version: 0.2
phases:
pre_build:
commands:
- echo Logging in to Amazon ECR...
- aws ecr get-login-password --region eu-central-1 | docker login --username AWS --password-stdin REGISTRY_URI
- COMMIT_HASH=$(echo $CODEBUILD_RESOLVED_SOURCE_VERSION | cut -c 1-7)
- IMAGE_TAG=${COMMIT_HASH:=latest}
- REPOSITORY_URI=REGISTRY_URI/REPOSITORY_NAME
build:
commands:
- echo Build started on `date`
- echo Building the Docker image...
- docker build -t $REPOSITORY_URI:latest .
- docker tag $REPOSITORY_URI:latest $REPOSITORY_URI:$IMAGE_TAG
- apt-get install jq -y
post_build:
commands:
- echo Build completed on `date`
- echo pushing to repo
- docker push $REPOSITORY_URI:latest
- docker push $REPOSITORY_URI:$IMAGE_TAGThe deployment part
I chose ECS as a deployment provider and went through the configuration steps. I chose the cluster and the service and then I had to provide the imagedefinitions.json file. I was not sure what to put there so I decided to leave it empty and see what happens. The pipeline was created and I ran it. The build stage was successful but the deployment stage failed with an error saying that the imagedefinitions.json file is missing. I figured it out that I have to provide a file named imagedefinitions.json that contains the image name and tag, because CodePipeline deploy worker uses this file to know where to pull the image from in order to deploy it to ECS. I found this article helpful to understand the imagedefinitions.json file format. The article documenet the JSON structure of the file and also provides an example of the file content. The content of the file should look like the following snippet, and it should be stored as artifact in the build stage.[
{
"name": "sample-app",
"imageUri": "IMAGE_URI"
}
]docker push $REPOSITORY_URI:$IMAGE_TAG - ContainerName="CONTAINER_NAME"
- ImageURI=$(cat imageDetail.json | jq -r '.ImageURI')
- printf '[{"name":"CONTAINER_NAME","imageUri":"IMAGE_URI"}]' > imagedefinitions.json
- sed -i -e "s|CONTAINER_NAME|$ContainerName|g" imagedefinitions.json
- sed -i -e "s|IMAGE_URI|$ImageURI|g" imagedefinitions.json
- cat imagedefinitions.jsonartifacts:
files:
- imagedefinitions.jsonAs a conclusion, The imagedefinitions.json file is required when using CodePipeline to deploy to ECS. It should be stored as an artifact in the build stage and it should contain the image name and tag. The file can be generated using the commands in the post_build phase of the buildspec.yml file.
🤟 Also, I'm available for consulting and freelance work. If you need help with your AWS infrastructure, feel free to reach out to me at helloismajl@gmail.com
References
https://docs.aws.amazon.com/codepipeline/latest/userguide/file-reference.html#pipelines-create-image-definitionshttps://stackoverflow.com/questions/58849736/did-not-find-the-image-definition-file-imagedefinitions-json
https://docs.aws.amazon.com/codebuild/latest/userguide/sample-docker.html