terraform-aws-jenkins

An AWS EC2 setup running Jenkins provisioned via Terraform.

Requirements

• Terraform 0.9+

Prerequisites

• Please ensure you have the aws cli application working

• Before you run this repo, ensure you've ran the terraform-aws-init repo. That will ensure you have the terraform-states s3 bucket and a root-ssh-key for shell access to jenkins.

Instructions

The scripts can be called with the following parameters:

tf-plan.sh <s3_prefix> <env_name> <region> <availability_zones> <ssh_key_name>

You must pass in the s3prefix you used in the terraform-aws-init s3prefix to this script. Ex. ./tf-plan.sh optimus-prime-87

Once the plan runs and everything looks good, you can provision your jenkins instance by running ./tf-apply.sh <s3_prefix>

That is all there is to it! You can then see the IP of your Jenkins instance and ssh or browse to it.

To ssh ssh [email protected]

Jenkins UI can be seen browsed at x.x.x.x:8080 where x.x.x.x is the output IP address of the ec2 instance.

PLEASE SECURE JENKINS ASAP BY PUTTING INTO A PRIVATE SUBNET OR ADDING PASSWORD AUTH

Post Instructions

After Jenkins is setup, you'll need to start editing configs in your forked terraform-example-manifest repo that you created earlier.

SSH into your Jenkins instance and do the following:

Setup GitHub Credentials with Jenkins

• SSH into the Jenkins EC2 and switch to the Jenkins user. Move to the Jenkins home directory.

  sudo su jenkins && cd ~/

• Run the git config script

  ./configure-git.sh

• Copy/Paste your key into GitHub and Save *

• Clone something from GitHub in the CLI on the Jenkins server and accept the fingerprint.

Setup Docker Credentials with Jenkins

Instructions when using your own DockerHub account:

• Login to your Docker account via CLI

  docker login

• Copy the token from the ~./docker/config.json file

• In your forked terraform-example-manifest repo, edit the docker.json file.

  • Update dockerhub_email with your docker hub email.

  • Update dockerhub_token with the token you just copied from the config.json file.

Instructions when using premade DockerHub account

• Make sure you switch to the Jenkins user.

sudo su jenkins && cd ~/

• Create the “.docker” directory

mkdir ~/.docker

• Create a ~./docker/config.json file

vim ~/.docker/config.json

• Paste the following into this file:

{
  "auths": {
    "https://index.docker.io/v1/": {
      "auth": "dGhpc2lzanVzdGF0ZXN0OnRlY2h0djEyMQ=="
    }
  }
}

Your manifest (config) repo for your environment variables

• Fork the terraform-example-manifest repo and change the terraform.json file with your s3 prefix. Push that change up and copy the github ssh clone url.

• Head over to your Jenkins instance in a web browser and click Manage Jenkins -> Configure System -> Check off Environment Variables and add one called MANIFEST_REPO and paste your manifest url you copied earlier here. Ex. [email protected]:OGProgrammer/terraform-example-manifest.git

Setup Jenkins Jobs

Follow from slide 36 and on for how to setup the rest of the dev/prod jobs.

Google Slides can be found here

Important Notes

If you provision this Jenkins terraform script with something that deletes and creates a new Jenkins instance, (because yes that can happen) don't freak out. You can login to the AWS EC2->Volumes page and detach the new drive and add the old one back. All your Jenkins/Docker data should be stored on this volume.

1. SSH into Jenkins and stop Jenkins sudo service jenkins stop
2. Unmount the new drive sudo umount /dev/xvdf
3. Detach the volume in the AWS dashboard.

Delete drives not in use anymore as they will pile up even after a terraform destroy.

Destroying

Just run ./tf-destroy.sh <s3_prefix> but in order to finalize cleanup you'll need to delete one last thing manually.

There is a 250 GB drive mounted to Jenkins that is not destroyed when you destroy the instance with terraform, you have to manually goto AWS EC2 Volumes and destroy the drive that way.

This is to prevent critical production data from getting destroyed but yea, I even keep forgetting to kill these drives...

TODOs

• This could probably be further improved by using Ansible, Puppet, Chef, or something like that to provision any OS. This is really locked to Debian at the moment.

• Have a way to automatically create the jenkins jobs based off a manifest in s3.

• Combine jobs for less work (plan and apply in one) but more risk unless you add an approval step.

• Refactor groovy scripts to look for credentials for docker and github rather than going onto box to configure.

Built & Maintained by @OGProgrammer

Support Your Local User Groups
http://php.ug/

Check out our PHP UG in Las Vegas, NV
http://PHPVegas.com

Support your local tech scene!
#VegasTech

Share your knowledge!
Become a speaker, co-organizer, at your local user groups.
Contribute to your favorite open source packages (even if its a README ;)

Thank you! ☺

-Josh

Paid support and training services available at http://RemoteDevForce.com

Slides: https://docs.google.com/presentation/d/1KeZn1z-p2zWoeeI8hxI-B7DI1wkBDjSjXU5_k1OsjJM/edit?usp=sharing