Overview

Project to build a base ubuntu image for Sage Bionetworks science use and deploy it to AWS. The image is based on the latest Ubuntu Bionic release available at the point of build and also includes the following:

• Python3 and supporting libs
• Docker binaries
• AWS CLI client

Workflow

The workflow to provision AWS AMI is done using pull requests. Just make changes with PRs and when PR is merged a packer build will kick off which builds the image and deploys it to AWS.

Note: The image will automatically be named gitrepo-branch (i.e. myrepo-master)

Requirements

• Install packer (Use this shell script)

Manual AMI Build

If you would like to test building an AMI run:

cd src
export IMAGE_NAME=<repo name>-test
packer build -var AwsProfile=my-aws-account -var AwsRegion=us-east-1 -var ImageName=${IMAGE_NAME} template.json

Packer will do the following:

• Create a temporary EC2 instance, configure it with shell/ansible/puppet/etc. scripts.
• Create an AMI from the EC2
• Delete the EC2

Note: Packer deploys a new AMI to the AWS account specified by the AwsProfile

Deploy a release

To build and deploy a release image just git tag the repo and a packer build will kick off which builds the image and deploys it to AWS.

Note: The image will automatically be named gitrepo-tag (i.e. myrepo-v1.0.0)

Contributions

Contributions are welcome.

Requirements:

• Install pre-commit app
• Clone this repo
• Run pre-commit install to install the git hook.

Testing

As a pre-deployment step we syntatically validate our packer json files with pre-commit.

Please install pre-commit, once installed the file validations will automatically run on every commit. Alternatively you can manually execute the validations by running pre-commit run --all-files.

Deployments

Travis runs packer which temporarily deploys an EC2 to create an AMI.

Continuous Integration

We have configured Travis to deploy cloudformation template updates.

Possible Upcoming changes

Replace AWS CLI client binaries with docker container version and make system-wide shell alias

Issues

• https://sagebionetworks.jira.com/projects/IT

Secrets

• We use the AWS SSM to store secrets for this project. Sceptre retrieves the secrets using a sceptre ssm resolver and passes them to the cloudformation stack on deployment.