Infrastructure as Code (IAC) Cookbook: Automate complex infrastructures

To step through this recipe, you will need the following:

Open a terminal and type the following code:

$ vagrant box add ubuntu/xenial64
==> box: Loading metadata for box 'ubuntu/xenial64'
    box: URL: https://atlas.hashicorp.com/ubuntu/xenial64
==> box: Adding box 'ubuntu/xenial64' (v20160815.0.0) for provider: virtualbox
    box: Downloading: https://atlas.hashicorp.com/ubuntu/boxes/xenial64/versions/20160815.0.0/providers/virtualbox.box
==> box: Successfully added box 'ubuntu/xenial64' (v20160815.0.0) for 'virtualbox'!

Vagrant knows where to look for the latest version for the requested box on the Atlas service and automatically downloads it over the Internet. All boxes are stored by default in ~/.vagrant.d/boxes.

If you're interested in creating your own base Vagrant boxes, refer to Packer (https://www.packer.io/) and the Chef Bento project (http://chef.github.io/bento/).

To step through this recipe, you will need the following:

When you started Vagrant, it read the Vagrantfile, asking for a specific box to run (Ubuntu Xenial). If you previously added it, it will launch it right away through the default hypervisor (in this case, VirtualBox), or if it's a new box, download it for you automatically. It created the required virtual network interfaces, then the Ubuntu VM got a private IP address. Vagrant took care of configuring SSH by exposing an available port and inserting a default key, so you can log into it via SSH without problems.

To step through this recipe, you will need the following:

Follow these steps to enable VirtualBox Guest Additions in Vagrant:

Vagrant plugins are automatically installed from the vendor's website, and made available globally on your system for all other Vagrant environments you'll run. Once the virtual machine is ready, the plugin will detect the operating system, decide if the Guest Additions need to be installed or not, and if they do, install the necessary tools to do that (compilers, kernel headers, and libraries), and finally download and install the corresponding Guest Additions.

Using Vagrant plugins also extends what you can do with the Vagrant CLI. In the case of the VirtualBox Guest Addition plugin, you can do a lot of things such as status checks, manage the installation, and much more:

$ vagrant vbguest --status
[default] GuestAdditions 5.0.26 running --- OK.

The plugin can later be called through Vagrant directly; here it's triggering the Guest Additions installation in the virtual machine:

$ vagrant vbguest --do install

To step through this recipe, you will need the following:

The Chef Bento project provides various multiprovider images we can use. For example, let's use a CentOS 7.2 with Vagrant (bento/centos-7.2) with this simplest Vagrantfile:

Vagrant.configure("2") do |config|
  config.vm.box = "bento/centos-7.2"
end

Start your CentOS 7.2 virtual environment and specify the hypervisor you want to run:

$ vagrant up --provider=vmware_fusion
$ vagrant ssh

You're now running a CentOS 7.2 Vagrant box using VMware!

Vagrant is powered by plugins extending its usage and capabilities. In this case, the Vagrant plugin for VMware delegates all the virtualization features to the VMware installation, removing the need for VirtualBox.

If VMware is your primary hypervisor, you'll soon be tired to always specify the provider in the command line. By setting the VAGRANT_DEFAULT_PROVIDER environment variable to the corresponding plugin, you will never have to specify the provider again, VMware will be the default:

$ export VAGRANT_DEFAULT_PROVIDER=vmware_fusion
$ vagrant up

To step through this recipe, you will need the following:

The VMware provider can be configured inside the following configuration blocks:

# VMware Fusion configuration
config.vm.provider "vmware_fusion" do |vmware|
  # enter all the vmware configuration here
end

# VMware Workstation configuration
config.vm.provider "vmware_workstation" do |vmware|
  # enter all the vmware configuration here
end

If the configuration is the same, you'll end up with a lot of duplicated code. Take advantage of the Ruby nature of the Vagrantfile and use a simple loop to iterate through both values:

["vmware_fusion", "vmware_workstation"].each do |vmware|
  config.vm.provider vmware do |v|
    # enter all the vmware configuration here
  end
end

Our default Bento CentOS 7.2 image has only 512 MB of RAM and one CPU. Let's double that for better performance using the vmx["numvcpus"] and vmx["memsize"] keys:

  ["vmware_fusion", "vmware_workstation"].each do |vmware|
    config.vm.provider vmware do |v|
      v.vmx["numvcpus"] = "2"
      v.vmx["memsize"] = "1024"
    end
  end

Start or restart your Vagrant machine to apply the changes:

$ vagrant up
[…]

Your box is now using two CPUs and 1 GB of RAM.

Virtual machine configuration is the last thing done by Vagrant before starting up. Here, it just tells VMware to allocate two CPUs and 1 GB of RAM to the virtual machine it's launching the way you would have done manually from inside the software.

Vagrant's authors may merge both plugins into one at some point in the future. The current 4.x version of the plugins is still split.

The VMX format is not very well documented by VMware. The possible keys and values can be found on most VMware Inc. documentation about VMX configuration.

To step through this recipe, you will need the following:

Some Vagrant boxes are available for multiple hypervisors, such as the CentOS 7 Bento box we previously used. This way, we can simply choose which one to use.

Let's start with our previous Vagrantfile including customizations for VMware:

Vagrant.configure("2") do |config|
  config.vm.box = "bento/centos-7.2"
  ["vmware_fusion", "vmware_workstation"].each do |vmware|
    config.vm.provider vmware do |v|
      v.vmx["numvcpus"] = "2"
      v.vmx["memsize"] = "1024"
    end
  end
end

How would we add the same configuration on VirtualBox as we have on VMware? Here's how to customize VirtualBox similarly in the Vagrantfile:

  config.vm.provider :virtualbox do |vb|
    vb.memory = "1024"
    vb.cpus = "2"
  end

Add this to your current Vagrantfile, reload and you'll get the requested resources from your hypervisor, be it VMware or VirtualBox.

It's nice, but we're still repeating ourselves with the values, leading to possible errors, omissions, or mistakes in the future. Let's take advantage once again of the Ruby nature of our Vagrantfile and declare some meaningful variables at the top of our file:

vm_memory = 1024
vm_cpus = 2

Now replace the four values by their variable names and you're done: you're centrally managing characteristics of the Vagrant environment you're using and distributing, whatever hypervisor you're using.

The simple fact that the Vagrantfile is a pure Ruby file helps creating powerful and dynamic configuration, by simply setting variables that we use later for all the providers.

To step through this recipe, you will need the following:

Here are some possible customizations for your Vagrant Virtual Machine.

config.vm.hostname = "vagrant-lab-1"

config.vm.box_check_update = false

config.vm.box_version = "2.2.9"

config.vm.post_up_message = "Use \"vagrant ssh\" to log into the box. This VM uses #{vm_cpus} CPUs and #{vm_memory}MB of RAM."

Vagrant.require_version ">= 1.8.0"

To step through this recipe, you will need the following:

We'll see how to use, access, and manipulate an NGINX container in Vagrant using Docker as a provider.

Vagrant.configure("2") do |config|
  config.vm.hostname = "vagrant-docker-1"
  config.vm.post_up_message = "HTTP access: http://localhost/"
  config.vm.provider "docker" do |docker|
      docker.image = "nginx:stable"
  end
end

$ vagrant up --provider=docker
Bringing machine 'default' up with 'docker' provider...
==> default: Creating the container...
[…]
==> default: HTTP access: http://localhost/

ENV['VAGRANT_DEFAULT_PROVIDER'] = 'docker'

  config.vm.provider "docker" do |docker|
      docker.image = "nginx:stable"
      docker.ports = ['80:80']
  end

$ curl http://localhost/

$ mkdir src; echo "<h1>Hello from Docker via Vagrant<h1>" > src/index.html

  config.vm.provider "docker" do |docker|
      docker.image = "nginx:stable"
      docker.ports = ['80:80']
      docker.volumes = ["#{Dir.pwd}/src:/usr/share/nginx/html"]
  end

$ curl http://localhost
<h1>Hello from Docker via Vagrant<h1>

$ sudo setenforce 0

You can choose not to use your local or default Docker installation, but instead use a dedicated VM, maybe to reflect production or a specific OS (such as CoreOS). In this case, you can specify a dedicated Vagrantfile as follows:

config.vm.provider "docker" do |docker|
docker.vagrant_vagrantfile = "docker_host/Vagrantfile"
end

To step through this recipe, you will need the following:

The previous example allows only one container to be launched simultaneously, which is sad considering the power of Docker. Let's define multiple containers and start by creating a front container (our previous NGINX):

  config.vm.define "front" do |front|
    front.vm.provider "docker" do |docker|
      docker.image = "nginx:stable"
      docker.ports = ['80:80']
      docker.volumes = ["#{Dir.pwd}/src:/usr/share/nginx/html"]
    end
  end

Now how about creating an application container, maybe a blog engine such as Ghost? Ghost publishes a ready-to-use container on the Docker Hub, so let's use that (version 0.9.0 at the time of writing) and expose on TCP/8080 the application container listening on TCP/2368:

  config.vm.define "app" do |app|
    app.vm.provider "docker" do |docker|
      docker.image = "ghost:0.9.0"
      docker.ports = ['8080:2368']
    end
  end

Check if you can access the blog on http://localhost:8080 and NGINX on http://localhost:

$ curl -IL http://localhost:8080
HTTP/1.1 200 OK
X-Powered-By: Express
[…]

$ curl -IL http://localhost
HTTP/1.1 200 OK
Server: nginx/1.10.1

Now let's use NGINX for what it's for—serving the application. Configuring NGINX as a reverse proxy is beyond the scope of this book, so just use the following simple configuration for the nginx.conf file at the root of your working folder:

server {
  listen 80;
  location / {
    proxy_set_header   X-Real-IP $remote_addr;
    proxy_set_header   Host      $http_host;
    proxy_pass         http://app:2368;
  }
}

Change the configuration of the front container in Vagrant to use this configuration, remove the old index.html as we're not using it anymore, and link this container to the app container:

  config.vm.define "front" do |front|
    front.vm.provider "docker" do |docker|
      docker.image = "nginx:stable"
      docker.ports = ['80:80']
      docker.volumes = ["#{Dir.pwd}/nginx.conf:/etc/nginx/conf.d/default.conf"]
      docker.link("app:app")
    end
  end

Linking the app container makes it available to the front container, so now there's no need to expose the Ghost blog container directly, let's make it simpler and more secure behind the reverse proxy:

  config.vm.define "app" do |app|
    app.vm.provider "docker" do |docker|
      docker.name = "app"
      docker.image = "ghost:0.9.0"
    end
  end

We're close! But this setup will eventually fail for a simple reason: our systems are too fast, and Vagrant parallelizes the startup of virtual machines by default, and also does this for containers. Containers start so fast that the app container may not be ready for NGINX when it's started. To ensure sequential startup, use the VAGRANT_NO_PARALLEL environment variable at the top of the Vagrantfile:

ENV['VAGRANT_NO_PARALLEL'] = 'true'

Now you can browse to http://localhost/admin and start using your Ghost blog in a container, behind a NGINX reverse proxy container, with the whole thing managed by Vagrant!

You can access the containers logs directly using Vagrant:

$ vagrant docker-logs --follow
==> app: > ghost@0.9.0 start /usr/src/ghost
==> app: > node index
==> app: Migrations: Creating tables...
[…]
==> front: 172.17.0.1 - - [21/Aug/2016:10:55:08 +0000] "GET / HTTP/1.1" 200 1547 "-" "Mozilla/5.0 (X11; Fedora; Linux x86_64; rv:48.0) Gecko/20100101 Firefox/48.0" "-"
==> app: GET / 200 113.120 ms - -
[…]

version: '2'
services:
  front:
    image: nginx:stable
    volumes:
      - "./nginx.conf:/etc/nginx/conf.d/default.conf"
    restart: always
    ports:
      - "80:80"
    depends_on:
      - app
    links:
      - app
  app:
    image: ghost:0.9.0
    restart: always

To step through this recipe, you will need the following:

Begin by installing the plugin:

$ vagrant plugin install vagrant-aws

A requirement of this plugin is the presence of a dummy Vagrant box that does nothing:

$ vagrant box add dummy https://github.com/mitchellh/vagrant-aws/raw/master/dummy.box

Remember how we configured the Docker provider in the previous recipes? This is no different:

config.vm.provider :aws do |aws, override|
  # AWS Configuration
  override.vm.box = "dummy"
end

Then, defining an application VM will consist of specifying which provider it's using (AWS in our case), the Amazon Machine Image (AMI) (Ubuntu 16.04 LTS in our case), and a provisioning script that we creatively named script.sh.

You can find other AMI IDs at http://cloud-images.ubuntu.com/locator/ec2/:

config.vm.define "srv-1" do |config|
    config.vm.provider :aws do |aws|
      aws.ami = "ami-c06b1eb3"
    end
    config.vm.provision :shell, :path => "script.sh"
end

So what is the AWS-related information we need to fill in so Vagrant can launch servers on AWS?

We need the AWS Access Keys, preferably from environment variables so you don't hardcode them in your Vagrantfile:

aws.access_key_id = ENV['AWS_ACCESS_KEY_ID']
aws.secret_access_key = ENV['AWS_SECRET_ACCESS_KEY']

Indicate the region and availability zone where you want the instance to start:

aws.region = "eu-west-1"
aws.availability_zone = "eu-west-1a"

Include the instance type; here, we've chosen the one included in the AWS free tier plan so it won't cost you a dime with a new account:

aws.instance_type = "t2.micro"

Indicate in which security group this instance will live (it's up to you to adapt the requirements to your needs):

aws.security_groups = ['iac-lab']

Specify the AWS keypair name, and override the default SSH username and keys:

aws.keypair_name = "iac-lab"
override.ssh.username = "ubuntu"
override.ssh.private_key_path = "./keys/iac-lab.pem"

Under some circumstances, you can experience a bug with NFS while using Vagrant and AWS EC2, so I choose to disable this feature:

override.nfs.functional = false

Finally, it's a good practice to tag the instances, so you can later find out where they come from:

aws.tags = {
  'Name'   => 'Vagrant'
}

Add a simple shell script that will install Docker and docker-compose, then execute the docker-compose file:

#!/bin/sh
# install Docker
curl -sSL https://get.docker.com/ | sh
# add ubuntu user to docker group
sudo usermod -aG docker ubuntu
# install docker-compose
curl -L https://github.com/docker/compose/releases/download/1.8.0/docker-compose-`uname -s`-`uname -m` > /usr/local/bin/docker-compose
sudo chmod +x /usr/local/bin/docker-compose
# execute the docker compose file
cd /vagrant
docker-compose up -d

Include both NGINX configuration and docker-compose.yml files from the previous recipe and you're good to go:

$ vagrant up
Bringing machine 'srv-1' up with 'aws' provider...
[…]
==> srv-1: Launching an instance with the following settings...
==> srv-1:  -- Type: t2.micro
==> srv-1:  -- AMI: ami-c06b1eb3
==> srv-1:  -- Region: eu-west-1
[…]
==> srv-1: Waiting for SSH to become available...
==> srv-1: Machine is booted and ready for use!
[…]
==> srv-1:  docker version
[…]
==> srv-1: Server:
==> srv-1:  Version:      1.12.1
[…]
==> srv-1: Creating vagrant_app_1
==> srv-1: Creating vagrant_front_1

Open your browser at http://a.b.c.d/ (using the EC2 instance public IP) and you'll see your Ghost blog behind an NGINX reverse proxy, using Docker containers, using Vagrant on Amazon EC2.

A common usage for such a setup is for the developer to test the application in close to real production conditions, maybe to show a new feature to a remote product owner, replicate a bug seen only in this setup, or at some point in the CI. Once Docker containers have been built, smoke test them on EC2 before going any further.

To step through this recipe, you will need the following:

Here's how we would create one CentOS 7.2 machine with 512 MB of RAM and one CPU, in a private network with a fixed IP 192.168.50.11, and a simple shell output:

vm_memory = 512
vm_cpus = 1

Vagrant.configure("2") do |config|

  config.vm.box = "bento/centos-7.2"

  config.vm.provider :virtualbox do |vb|
    vb.memory = vm_memory
    vb.cpus = vm_cpus
  end

   config.vm.define "srv-1" do |config|
     config.vm.provision :shell, :inline => "ip addr | grep \"inet\" | awk '{print $2}'"
     config.vm.network "private_network", ip: "192.168.50.11", virtualbox__intnet: "true"
   end
end

To add a new machine to this network, we could simply duplicate the srv-1 machine definition, as in the following code:

config.vm.define "srv-2" do |config|
     config.vm.provision :shell, :inline => "ip addr | grep \"inet\" | awk '{print $2}'"
     config.vm.network "private_network", ip: "192.168.50.12", virtualbox__intnet: "true"
end

That's not very DRY, so let's take advantage of the Ruby nature of the Vagrantfile to create a loop that will dynamically and simply create as many virtual machines as we want.

First, declare a variable with the amount of virtual machines we want (2):

vm_num = 2

Then iterate through that value, so it can generate values for an IP and for a hostname:

(1..vm_num).each do |n|
    # a lan lab in the 192.168.50.0/24 range
    lan_ip = "192.168.50.#{n+10}"
    config.vm.define "srv-#{n}" do |config|
      config.vm.provision :shell, :inline => "ip addr | grep \"inet\" | awk '{print $2}'"
      config.vm.network "private_network", ip: lan_ip, virtualbox__intnet: "true"
    end
  end

This will create two virtual machines (srv-1 at 192.168.50.11 and srv-2 at 192.168.50.12) on the same internal network, so they can talk to each other.

Now you can simply change the value of vm_num and you'll easily spawn new virtual machines in seconds.

We can optionally go even further, using the following cloning and networking features.

config.vm.provider :virtualbox do |vb|
    vb.memory = vm_memory
    vb.cpus = vm_cpus
    vb.linked_clone = true
end

config.vm.network "private_network", ip: lan_ip, virtualbox__intnet: "iac-lab"

To step through this recipe, you will need the following:

Follow these steps for simulating a networked three-tier architecture app with Vagrant.

  config.vm.define "db-1" do |config|
    config.vm.hostname = "db-1"
    config.vm.network "private_network", ip: "10.30.0.11", virtualbox__intnet: "db_lan"
    config.vm.provision :shell, :inline => "sudo yum install -q -y epel-release"
    config.vm.provision :shell, :inline => "sudo yum install -q -y redis"
    config.vm.provision :shell, :inline => "sudo sed -i 's/bind 127.0.0.1/bind 127.0.0.1 10.30.0.11/' /etc/redis.conf"
    config.vm.provision :shell, :inline => "sudo systemctl enable redis"
    config.vm.provision :shell, :inline => "sudo systemctl start redis"
  end

#!/usr/bin/env node
var os = require("os");
var redis = require('redis');
var client = redis.createClient(6379, '10.30.0.11');
client.on('connect', function() {
    console.log('connected to redis on '+os.hostname()+' 10.30.0.11:6379');
});

var http = require('http');
http.createServer(function (req, res) {
  res.writeHead(200, {'Content-Type': 'text/plain'});
  res.end('Running on '+os.hostname()+'\n');
}).listen(8080);
console.log('HTTP server listening on :8080');

[Unit]
Description=Node App
After=network.target

[Service]
ExecStart=/srv/nodeapp/app.js
Restart=always
User=vagrant
Group=vagrant
Environment=PATH=/usr/bin
Environment=NODE_ENV=production
WorkingDirectory=/srv/nodeapp
[Install]
WantedBy=multi-user.target

# Tier 2: a scalable number of application servers
vm_app_num = 2
  (1..vm_app_num).each do |n|
    app_lan_ip = "10.20.0.#{n+10}"
    db_lan_ip = "10.30.0.#{n+100}"
    config.vm.define "app-#{n}" do |config|
      config.vm.hostname = "app-#{n}"
      config.vm.network "private_network", ip: app_lan_ip, virtualbox__intnet: "app_lan"
      config.vm.network "private_network", ip: db_lan_ip, virtualbox__intnet: "db_lan"
      config.vm.provision :shell, :inline => "sudo yum install -q -y epel-release"
      config.vm.provision :shell, :inline => "sudo yum install -q -y nodejs npm"
      config.vm.provision :shell, :inline => "sudo mkdir /srv/nodeapp"
      config.vm.provision :shell, :inline => "sudo cp /vagrant/app.js /src/nodeapp"
      config.vm.provision :shell, :inline => "sudo chown -R vagrant.vagrant /srv/"
      config.vm.provision :shell, :inline => "sudo chmod +x /srv/nodeapp/app.js"
      config.vm.provision :shell, :inline => "cd /srv/nodeapp; npm install redis"
      config.vm.provision :shell, :inline => "sudo cp /vagrant/nodeapp.service /etc/systemd/system"
      config.vm.provision :shell, :inline => "sudo systemctl daemon-reload"
      config.vm.provision :shell, :inline => "sudo systemctl start nodeapp"
    end
  end

events {
  worker_connections 1024;
}
http {
  upstream app {
    server 10.20.0.11:8080 max_fails=1 fail_timeout=1s;
    server 10.20.0.12:8080 max_fails=1 fail_timeout=1s;
  }
  server {
    listen 80;
    server_name  _;
    location / {
      proxy_set_header   X-Real-IP $remote_addr;
      proxy_set_header   Host      $http_host;
      proxy_pass         http://app;
    }
  }
}

  # Tier 1: an NGINX reverse proxy VM, available on http://localhost:8080
  config.vm.define "front-1" do |config|
    config.vm.hostname = "front-1"
    config.vm.network "private_network", ip: "10.10.0.11", virtualbox__intnet: "front_lan"
    config.vm.network "private_network", ip: "10.20.0.101", virtualbox__intnet: "app_lan"
    config.vm.network "forwarded_port", guest: 80, host: 8080
    config.vm.provision :shell, :inline => "sudo yum install -q -y epel-release"
    config.vm.provision :shell, :inline => "sudo yum install -q -y nginx"
    config.vm.provision :shell, :inline => "sudo cp /vagrant/nginx.conf /etc/nginx/nginx.conf"
    config.vm.provision :shell, :inline => "sudo systemctl enable nginx"
    config.vm.provision :shell, :inline => "sudo systemctl start nginx"
  end

To step through this recipe, you will need the following:

Let's start with the simplest Vagrant environment we know:

Vagrant.configure("2") do |config|
  config.vm.box = "bento/centos-7.2"
  config.vm.define "srv-1" do |config|
    config.vm.hostname = "srv-1"
  end
end

events {
  worker_connections 1024;
}
http {
  sendfile off;
  server {
    listen 80;
    server_name  _;
    root /srv/app/public ;
    try_files $uri $uri/ /index.php?q=$uri&$args;
    index index.php;
    location / {
      try_files $uri $uri/ /index.php?$query_string;
    }
    location ~ \.php$ {
      try_files $uri /index.php =404;
      fastcgi_split_path_info ^(.+\.php)(/.+)$;
      fastcgi_pass 127.0.0.1:9000;
      fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
      fastcgi_param PATH_INFO $fastcgi_script_name;
      include fastcgi_params;
    }
  }
}

config.vm.provision :shell, :path => "provision.sh"

config.vm.synced_folder "src/", "/srv/app"

config.vm.network "public_network", bridge: "en0: Wi-Fi (AirPort)"

==> srv-1: Available bridged network interfaces:
1) en0: Wi-Fi (AirPort)
[...]

cd /srv/app
composer create-project --prefer-dist laravel/laravel.

If you include the Vagrantfile directly with your project's code, coworkers or contributors will only have to run vagrant up to see it running.

Other Vagrantfile sharing options include Windows Sharing (smb), rsync (useful with remote virtual machines such as on AWS EC2), and even NFS.

A noticeable bug in the sharing feature using VirtualBox leads to corrupted or non-updating files. The workaround is to deactivate in the web server configuration sendfile, using NGINX:

sendfile off;

Using Apache, it is as follows:

EnableSendfile Off

To step through this recipe, you will need the following:

Let's start with this simple Vagrantfile:

Vagrant.configure("2") do |config|
  config.vm.box = "bento/centos-7.2"
  config.vm.define "blog" do |config|
    config.vm.hostname = "blog"
  end
end

We know we'll have to install some packages, so let's add a provisioning script to be executed:

    config.vm.provision :shell, :path => "provision.sh"

We'll want to hack locally on our Ghost blog, such as adding themes and more, so let's sync our src/ folder to the remote /srv/blog folder:

    config.vm.synced_folder "src/", "/srv/blog"

We want a local private network so we can access the virtual machine, with the 2368 TCP port (Ghost default) redirected to our host 8080 HTTP port:

    config.vm.network "private_network", type: "dhcp"
    config.vm.network "forwarded_port", guest: 2368, host: 8080

vagrant ssh
cd /srv/blog && sudo npm start --production
[…]
Ghost is running in production...
Your blog is now available on http://my-ghost-blog.com
Ctrl+C to shut down

server: {
            host: '0.0.0.0',
            port: '2368'
        }

cd /srv/blog && sudo npm start --production

$ vagrant share
==> srv-1: Detecting network information for machine...
[...]
==> srv-1: Your Vagrant Share is running! Name: anxious-cougar-6317
==> srv-1: URL: http://anxious-cougar-6317.vagrantshare.com

$ vagrant share --ssh
==> srv-1: Detecting network information for machine...
[...]
srv-1: Please enter a password to encrypt the key:
    srv-1: Repeat the password to confirm:
[...]
==> srv-1: You're sharing with SSH access. This means that another user
==> srv-1: simply has to run `vagrant connect --ssh subtle-platypus-4976`
==> srv-1: to SSH to your Vagrant machine.
[...]

$ vagrant connect --ssh subtle-platypus-4976
Loading share 'subtle-platypus-4976'...
[...]
[vagrant@srv-1 ~]$ head -n1 /srv/blog/config.js
// # Ghost Configuration

To step through this recipe, you will need the following:

Let's start with a minimal virtual machine named prod that simply boots a CentOS 7.2, like we have in our production environment:

Vagrant.configure("2") do |config|
  config.vm.box = "bento/centos-7.2"
  config.vm.define "prod" do |config|
    config.vm.hostname = "prod"
    config.vm.network "private_network", type: "dhcp"
  end

end

$ vagrant plugin install vagrant-omnibus
Installing the 'vagrant-omnibus' plugin. This can take a few minutes...
Installed the plugin 'vagrant-omnibus (1.4.1)'!

config.omnibus.chef_version = :latest

config.omnibus.chef_version = "11.18.12"

%w(vagrant-vbguest vagrant-omnibus).each do |plugin|
  unless Vagrant.has_plugin?(plugin)
    raise "#{plugin} plugin is not installed! Please install it using `vagrant plugin install #{plugin}`"
  end
end

package "httpd"

service "httpd" do
  action [ :enable, :start ]
end

    config.vm.provision :chef_solo do |chef|
      chef.add_recipe 'apache2'
    end

$ chef-solo --version
Chef: 11.18.12

$ httpd -v
Server version: Apache/2.4.6 (CentOS)

$ pidof httpd
13029 13028 13027 13026 13025 13024

  config.vm.define "staging" do |config|
    config.vm.hostname = "staging"
    config.omnibus.chef_version = "12.13.37"
    config.vm.network "private_network", type: "dhcp"
    config.vm.provision :chef_solo do |chef|
      chef.add_recipe 'apache2'
    end
  end

$ vagrant ssh staging
$ chef-solo --version
Chef: 12.13.37
$ httpd -v
Server version: Apache/2.4.6 (CentOS)
$ pidof httpd
13029 13028 13027 13026 13025 13024

Here are more ways of controlling a Vagrant environment, and use even better Chef tooling inside it.

config.vm.define "prod", primary: true do |config|
  […]
end

config.vm.define "staging", autostart: false do |config|
  […]
end

To step through this recipe, you will need the following:

Because writing complex Ansible playbooks is out of the scope of this book, we'll use a very simple one, so you can learn more about Ansible later and still reuse this recipe.

---
- hosts: all
  become: yes
  tasks:
    - name: Enable EPEL
      yum: name=epel-release state=present
    - name: Create a Docker group
      group: name=docker state=present
    - name: Add the vagrant user to Docker group
      user: name=vagrant groups=docker append=yes
    - name: Install Docker
      yum: name=docker state=present
    - name: Enable and Start Docker Daemon
      service: name=docker state=started enabled=yes

Vagrant.configure("2") do |config|
  config.vm.box = "bento/centos-7.2"
  config.vm.define "srv-1" do |config|
    config.vm.hostname = "srv-1"
    config.vm.network "private_network", type: "dhcp"
  end
end

    config.vm.provision "ansible" do |ansible|
      ansible.playbook = "playbook.yml"
    end

$ vagrant ssh
[vagrant@srv-1 ~]$ systemctl status docker
[vagrant@srv-1 ~]$ docker --version
Docker version 1.10.3, build d381c64-unsupported
[vagrant@srv-1 ~]$ docker run -it --rm alpine /bin/hostname
0f44a4d7afcd

What if for some reason you don't or can't have Ansible installed on your host machine? Alternatively, maybe you need a specific Ansible version on your Vagrant box to mimic production and you don't want to mess with your local Ansible installation. There's an interesting variant Ansible provider you can use: it will either use Ansible directly from the guest VM, and if it's not installed, it will install it from official repositories or PIP. You can use this very simple default configuration:

    config.vm.provision "ansible_local" do |ansible|
      ansible.playbook = "playbook.yml"
    end

You can also use the following command:

$ vagrant up
[…]
==> srv-1: Running provisioner: ansible_local...
    srv-1: Installing Ansible...
    srv-1: Running ansible-playbook...
[…]

Log in to the box via SSH and check that Ansible is locally installed with the latest version:

$ vagrant ssh
$ ansible --version
ansible 2.1.1.0

If your use case is different, you can use more precise deployment options, to be able to fix an Ansible version number using PIP (here, version 1.9.6 instead of the latest 2.x series):

    config.vm.provision "ansible_local" do |ansible|
      ansible.version = "1.9.6"
      ansible.install_mode = :pip
      ansible.playbook = "playbook.yml"
    end

You can also use the following command:

$ vagrant up
[…]
==> srv-1: Running provisioner: ansible_local...
    srv-1: Installing Ansible...
    srv-1: Installing pip... (for Ansible installation)
    srv-1: Running ansible-playbook...

Inside the Vagrant guest, you can now check for the PIP and Ansible versions:

$ pip --version
pip 8.1.2 from /usr/lib/python2.7/site-packages (python 2.7)
$ ansible --version
ansible 1.9.6

You can also check if our playbook has been installed correctly with the old 1.x Ansible version:

$ docker version

Also check if Docker is installed, and verify now it's working as the Vagrant user:

$ docker run -it --rm alpine ping -c2 google.com
PING google.com (216.58.211.78): 56 data bytes
64 bytes from 216.58.211.78: seq=0 ttl=61 time=22.078 ms
64 bytes from 216.58.211.78: seq=1 ttl=61 time=21.061 ms

To step through this recipe, you will need the following:

CoreOS doesn't host its official images at the default location on Atlas, it hosts it itself. So, we have to specify the full URL to the Vagrant box in our Vagrantfile:

Vagrant.configure("2") do |config|
  config.vm.box = https://stable.release.core-os.net/amd64-usr/current/coreos_production_vagrant.box
end

As CoreOS is a minimal OS, it doesn't support any of the VirtualBox guest addition tools, so we'll disable them, and don't try anything if we (most likely) have the vagrant-vbguest plugin:

  config.vm.provider :virtualbox do |vb|
      vb.check_guest_additions = false
      vb.functional_vboxsf     = false
  end

  if Vagrant.has_plugin?("vagrant-vbguest") then
    config.vbguest.auto_update = false
  end

Let's create a new VM definition, using the CoreOS Vagrant box:

  config.vm.define "core-1" do |config|
    config.vm.hostname = "core-1"
    config.vm.network "private_network", type: "dhcp" 
  end

We now need to run the mariadb and wordpress official containers from the Docker Hub. Using Docker directly, we would have run the following:

$ docker run -d --name mariadb -e MYSQL_ROOT_PASSWORD=h4ckm3 mariadb
$ docker run -d -e WORDPRESS_DB_HOST=mariadb -e 'WORDPRESS_DB_PASSWORD=h4ckm3 --link mariadb:mariadb -p 80:80 wordpress

Let's translate this into our Vagrantfile:

db_root_password = "h4ckm3"
config.vm.provision "docker" do |docker|
      docker.run "mariadb",
        args: "--name 'mariadb' -e 'MYSQL_ROOT_PASSWORD=#{db_root_password}'"
      docker.run "wordpress",
        args: "-e 'WORDPRESS_DB_HOST=mariadb' -e 'WORDPRESS_DB_PASSWORD=#{db_root_password}' --link 'mariadb:mariadb' -p '80:80'"
    end

Vagrant this up ($ vagrant up), and you'll access a ready-to-use WordPress installation running on CoreOS:

$ curl -IL http://172.28.128.3/wp-admin/install.php
HTTP/1.1 200 OK
Date: Thu, 25 Aug 2016 10:54:17 GMT
Server: Apache/2.4.10 (Debian)
X-Powered-By: PHP/5.6.25
Expires: Wed, 11 Jan 1984 05:00:00 GMT
Cache-Control: no-cache, must-revalidate, max-age=0
Content-Type: text/html; charset=utf-8

The CoreOS team proposes a full Vagrant environment to try and manipulate a CoreOS cluster https://github.com/coreos/coreos-vagrant. You'll then be able to try all CoreOS features and configuration options for all release channels (alpha, beta, or stable).

Other operating systems such as Ubuntu or CentOS are fully supported to provision Docker containers, even if Docker isn't installed at first on the base image. Vagrant will install Docker for you, so it will work transparently and run the containers as soon as it's installed.