Sepia Deployment Scripts

Scripts to setup a continuous deployment pipeline on AWS.

Prerequisites

• Dockerhub account
• AWS account
• Github Account
• Travis Account

Initial setup

Sample repositories that can be used to set up a sample pipeline:

Fork these repositories:

• Sample Liferay application
• Deployment definitions for the sample application

Then, add your forked liferay-game repository to your Travis account and make sure you set the following environment variables (keep them not visible):

Variable	Definition
AWS_ACCOUNT_ID	The ACCOUNT ID of your AWS account
AWS_ACCESS_KEY_ID	Your AWS ACCESS KEY ID
AWS_SECRET_ACCESS_KEY	Your AWS SECRET ACCESS KEY
DOCKER_ORG	Your public organization in Dockerhub
DOCKER_USER	Your Dockerhub user
DOCKER_PWD	Your Dockerhub password
DOCKER_AUTH_TOKEN	Your Docker Auth Token
GITHUB_TOKEN	A Github personal access token with repo access to your liferay-game-deployment repository
GITHUB_USER	The Github user of your liferay-game-deployment repository

To obtain the value for DOCKER_AUTH_TOKEN follow the steps described in Using Images from a Private Repository. The value you need to enter is referred on that link as auth_token. If you are running Docker on MAC, in order to get the right value of the token make sure you have configure Docker NOT to store docker logins in macOS keychain. To do so, got to Docker > Preferences and uncheck the option "Securely store docker logins in macOS keychain". On a terminal execute:

docker logout
docker login -u<DOCKER_USER> -p<DOCKER_PWD>

The value of the DOCKER_AUTH_TOKEN can be extracted from ~/.docker/config.json

{
  "auths" : {
    "https://index.docker.io/v1/" : {
      "auth" : "auth_token"
    }
  }
}

To obtain the value for GIHUB_TOKEN follow these steps: Creating a token.

Creating and updating the pipeline

In Travis, launch a first build of your liferay-game repository. Apart from running the tests, it will automatically build and publish a Docker image to your Dockerhub organization. Your liferay-game-deployment repository will be updated with the version of the recently generated Docker image. Then, all the elements of your continuous delivery pipeline will be automatically generated in your AWS account. Finally, deployments and tests of your Docker image will execute on each environment of your pipeline, until it is ready for production.

Now perform any change in the code of the liferay-game and push it to your Github master branch. The same process is repeated, except for the generation of the delivery pipeline in AWS, which is skipped because it was created during the first build.

Creating and updating a cloudformation stack

To create or update a stack just execute the command

aws/deployment-cloudformation/create-or-update-stack.sh -c CONFIG_DIR

where CONFIG_DIR is the absolute path to a directory that should follow the structure:

   - CONFIG_DIR
     - params
          *.json
     - templates
          template.json

where the json files located under the params directory should follow the structure:

[
   {
     "ParameterKey": "string",
     "ParameterValue": "string",
     "UsePreviousValue": true|false,
     "ResolvedValue": "string"
   }
   ...
 ]

A new stack will be created for every params/*.json file with the name stack-CONFIG_DIR_NAME-JSON_FILE_NAME and the content of the templates/template.json file.

To know more about to how write a AWS stack, please read the AWS Cloudformation documentation