j-r-j / aws_training_elixirconf_2019

Training Files for Elixir Conf 2019 workshop Elixir In the Jungle

Elixir In The Jungle Supplemental Files

These files are for educational purposes and take some liberties with configuration and security for the sake of brevity.

Their purpose is as a learning platform, and not as a solution unto themselves.

• Navigation
• General Conventions
• Challenge Solutions

Navigation

Each subfolder contains the slides and files specific to that part of the training.

• Part 1: The Physical Machine
• Part 2: The VPC and Bastion, as well as the example Phoenix app configured for CodeDeploy
• Part 3: The Pipeline, an inital ASG, Launch Configuration, CodeDeploy Application etc
• Part 4: An example Phoenix application. Terraform files can be found in the ops directory, and important runtime scripts at scripts. The appspec.yml file contains the CodeDeploy setup.

Notable readme files:

• 2_building_in_the_cloud/exercise/README.md
• 3_a_deployment_pipeline/exercise/README.md
• 4_the_details/exercise/web_app/ops/README.md

General Conventions

The .env file

2_building_in_the_cloud/exercise/.env.sample containes a sample .env file used for setting AWS credentials.

You can use it by copying to a .env file cp .env.sample .env and updating it with your own credentials.

You can then source the file source .env before running Terraform commands to add your credentials to the environment.

Terraform Shared State

** Note **

The Bucket used for terraform state is created and maintained outside of Terraform.

It will need to have a unique name, and be changed in each main.tf file for your use case.

Some variables in Part 3, are dependent on state/outputs from Part 2.

This is done via the terraform_remote_state data block found in the 3_a_deployment_pipeline/exercise/part_3_the_pipeline/main.tf file:

data "terraform_remote_state" "vpc" {
  backend = "s3"
  config = {
    bucket = "elixir-in-the-jungle-b"
    key = "vpc_state"
    region = "us-east-2"
  }

  workspace = "${terraform.workspace}"
}

Variables that are output from the VPC remote state can then be used like: data.terraform_remote_state.vpc.outputs.public_subnet_id

Terraform workspaces

Terraform has the concept of workspaces, these can be used to separate stacks like in the case of staging environments.

If you wish to spin up multiple environments there are a few things that need to change.

Most of the changes are in the variables.tf files, eg if you're adding an environment "staging" to the example files.

variable "vpc_cidr_blocks" {
  default = {
    example = "10.20.0.0/16"
  }
}

should change to something like

variable "vpc_cidr_blocks" {
  default = {
    example = "10.20.0.0/16"
    staging = "10.30.0.0/16"
  }
}

Additionally a new .env.example in the 3_a_deployment_pipeline/exercise/part_3_the_pipeline/ folder using the convention .env.YOUR_ENVIRONMENT

This file is used to provision the environment varibles for the app servers in that environment.

How to change regions

There are a few placed the region has to be hard coded.

In each part's main.tf file there will be remote state blocks referencing the remote state bucket and its region.

terraform {
  backend "s3" {
    bucket = "elixir-in-the-jungle-b"
    key = "app_servers_state"
    region = "us-east-2"
  }
}

data "terraform_remote_state" "vpc" {
  backend = "s3"
  config = {
    bucket = "elixir-in-the-jungle-b"
    key = "vpc_state"
    region = "us-east-2"
  }

  workspace = "${terraform.workspace}"
}

In addition there is one script (responsible for loading environment variables from S3) that has the region hard coded:

2_building_in_the_cloud/exercise/web_app/scripts/getenv and similarly 4_the_details/exercise/web_app/scripts/getenv

Challenge Solutions

Part 2 - the build script exercise/web_app/bin/build

ssh -T $1 <<-SSHCMDS
    mkdir -p builds &&
    tar -C builds -xf $_archive &&
    cd builds &&
    mix local.rebar --force &&
    mix local.hex --force &&
    mix deps.get --only prod &&
    cd assets &&
    npm install &&
    npm run deploy &&
    cd .. &&
    export MIX_ENV=prod &&
    mix phx.digest &&
    mix release --overwrite --path ./release &&
    cp -r ./scripts ./release/ &&
    cp ./appspec.yml ./release &&
    tar -C release . -zcf latest.tar.gz &&
    publish_release ./latest.tar.gz latest.tar.gz &&
    cd .. &&
    rm $_archive &&
    rm -rf builds
SSHCMDS

Part 3 - CodePipeline exercise/part_3_the_pipeline/codepipeline.tf

resource "aws_codepipeline" "app_servers" {
  name = "elixir-in-the-jungle-${terraform.workspace}"
  role_arn = "${aws_iam_role.codepipeline_role.arn}"

  artifact_store {
    location = "${data.terraform_remote_state.vpc.outputs.build_bucket_name}"
    type = "S3"
  }

  stage {
    name = "Source"

    action {
      name = "App-Change"
      category = "Source"
      owner = "AWS"
      provider = "S3"
      version = "1"
      output_artifacts = ["release"]

      configuration = {
        S3Bucket = "${data.terraform_remote_state.vpc.outputs.build_bucket_name}"
        S3ObjectKey = "latest.tar.gz"
        PollForSourceChanges = "true"
      }
    }

    action {
      name = "Config-Change"
      category = "Source"
      owner = "AWS"
      provider = "S3"
      version = "1"
      output_artifacts = ["env"]

      configuration = {
        S3Bucket = "${data.terraform_remote_state.vpc.outputs.build_bucket_name}"
        S3ObjectKey = "env"
        PollForSourceChanges = "true"
      }
    }

  }

  stage {
    name = "Deploy"

    action {
      name = "Deploy"
      category = "Deploy"
      owner = "AWS"
      provider = "CodeDeploy"
      input_artifacts = ["release"]
      version = "1"

      configuration = {
        ApplicationName = "${aws_codedeploy_app.app_servers.name}"
        DeploymentGroupName = "${aws_codedeploy_deployment_group.app_servers.deployment_group_name}"
      }
    }
  }
}