Vagrant AWS Provider


This is a Vagrant 1.2+ plugin that adds an AWS provider to Vagrant, allowing Vagrant to control and provision machines in EC2 and VPC.

NOTE: This plugin requires Vagrant 1.2+,

• Boot EC2 or VPC instances.
• SSH into the instances.
• Provision the instances with any built-in Vagrant provisioner.
• Minimal synced folder support via rsync.
• Define region-specific configurations so Vagrant can manage machines in multiple regions.
• Package running instances into new vagrant-aws friendly boxes

Install using standard Vagrant 1.1+ plugin installation methods. After installing, vagrant up and specify the aws provider. An example is shown below.

$ vagrant plugin install vagrant-aws
...
$ vagrant up --provider=aws
...

Of course prior to doing this, you'll need to obtain an AWS-compatible box file for Vagrant.

After installing the plugin (instructions above), the quickest way to get started is to actually use a dummy AWS box and specify all the details manually within a config.vm.provider block. So first, add the dummy box using any name you want:

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

And then make a Vagrantfile that looks like the following, filling in your information where necessary.

Vagrant.configure("2") do |config|
  config.vm.box = "dummy"

  config.vm.provider :aws do |aws, override|
    aws.access_key_id = "YOUR KEY"
    aws.secret_access_key = "YOUR SECRET KEY"
    aws.session_token = "SESSION TOKEN"
    aws.keypair_name = "KEYPAIR NAME"

    aws.ami = "ami-7747d01e"

    override.ssh.username = "ubuntu"
    override.ssh.private_key_path = "PATH TO YOUR PRIVATE KEY"
  end
end

And then run vagrant up --provider=aws.

This will start an Ubuntu 12.04 instance in the us-east-1 region within your account. And assuming your SSH information was filled in properly within your Vagrantfile, SSH and provisioning will work as well.

Note that normally a lot of this boilerplate is encoded within the box file, but the box file used for the quick start, the "dummy" box, has no preconfigured defaults.

If you have issues with SSH connecting, make sure that the instances are being launched with a security group that allows SSH access.

Note: if you don't configure aws.access_key_id or aws_secret_access_key it will attempt to read credentials from environment variables first and then from $HOME/.aws/. You can choose your AWS profile and files location by using aws.aws_profile and aws.aws_dir, however environment variables will always have precedence as defined by the AWS documentation. To use profile vagrantDev from your AWS files:

  # this first line can actually be omitted
  aws.aws_dir = ENV['HOME'] + "/.aws/"
  aws.aws_profile = "vagrantDev"

Every provider in Vagrant must introduce a custom box format. This provider introduces aws boxes. You can view an example box in the example_box/ directory. That directory also contains instructions on how to build a box.

The box format is basically just the required metadata.json file along with a Vagrantfile that does default settings for the provider-specific configuration for this provider.

This provider exposes quite a few provider-specific configuration options:

• access_key_id - The access key for accessing AWS
• ami - The AMI id to boot, such as "ami-12345678"
• availability_zone - The availability zone within the region to launch the instance. If nil, it will use the default set by Amazon.
• aws_profile - AWS profile in your config files. Defaults to default.
• aws_dir - AWS config and credentials location. Defaults to $HOME/.aws/.
• instance_ready_timeout - The number of seconds to wait for the instance to become "ready" in AWS. Defaults to 120 seconds.
• instance_check_interval - The number of seconds to wait to check the instance's state
• instance_package_timeout - The number of seconds to wait for the instance to be burnt into an AMI during packaging. Defaults to 600 seconds.
• instance_type - The type of instance, such as "m3.medium". The default value of this if not specified is "m3.medium". "m1.small" has been deprecated in "us-east-1" and "m3.medium" is the smallest instance type to support both paravirtualization and hvm AMIs
• keypair_name - The name of the keypair to use to bootstrap AMIs which support it.
• monitoring - Set to "true" to enable detailed monitoring.
• session_token - The session token provided by STS
• private_ip_address - The private IP address to assign to an instance within a VPC
• elastic_ip - Can be set to 'true', or to an existing Elastic IP address. If true, allocate a new Elastic IP address to the instance. If set to an existing Elastic IP address, assign the address to the instance.
• region - The region to start the instance in, such as "us-east-1"
• secret_access_key - The secret access key for accessing AWS
• security_groups - An array of security groups for the instance. If this instance will be launched in VPC, this must be a list of security group Name. For a nondefault VPC, you must use security group IDs instead (http://docs.aws.amazon.com/cli/latest/reference/ec2/run-instances.html).
• iam_instance_profile_arn - The Amazon resource name (ARN) of the IAM Instance Profile to associate with the instance
• iam_instance_profile_name - The name of the IAM Instance Profile to associate with the instance
• subnet_id - The subnet to boot the instance into, for VPC.
• associate_public_ip - If true, will associate a public IP address to an instance in a VPC.
• ssh_host_attribute - If :public_ip_address, :dns_name, or :private_ip_address, will use the public IP address, DNS name, or private IP address, respectively, to SSH to the instance. By default Vagrant uses the first of these (in this order) that is known. However, this can lead to connection issues if, e.g., you are assigning a public IP address but your security groups prevent public SSH access and require you to SSH in via the private IP address; specify :private_ip_address in this case.
• tenancy - When running in a VPC configure the tenancy of the instance. Supports 'default' and 'dedicated'.
• tags - A hash of tags to set on the machine.
• package_tags - A hash of tags to set on the ami generated during the package operation.
• use_iam_profile - If true, will use IAM profiles for credentials.
• block_device_mapping - Amazon EC2 Block Device Mapping Property
• elb - The ELB name to attach to the instance.
• unregister_elb_from_az - Removes the ELB from the AZ on removal of the last instance if true (default). In non default VPC this has to be false.
• terminate_on_shutdown - Indicates whether an instance stops or terminates when you initiate shutdown from the instance.
• endpoint - The endpoint URL for connecting to AWS (or an AWS-like service). Only required for non AWS clouds, such as eucalyptus.

These can be set like typical provider-specific configuration:

Vagrant.configure("2") do |config|
  # ... other stuff

  config.vm.provider :aws do |aws|
    aws.access_key_id = "foo"
    aws.secret_access_key = "bar"
  end
end

Note that you do not have to hard code your aws.access_key_id or aws.secret_access_key as they will be retrieved from the enviornment variables AWS_ACCESS_KEY and AWS_SECRET_KEY.

In addition to the above top-level configs, you can use the region_config method to specify region-specific overrides within your Vagrantfile. Note that the top-level region config must always be specified to choose which region you want to actually use, however. This looks like this:

Vagrant.configure("2") do |config|
  # ... other stuff

  config.vm.provider :aws do |aws|
    aws.access_key_id = "foo"
    aws.secret_access_key = "bar"
    aws.region = "us-east-1"

    # Simple region config
    aws.region_config "us-east-1", :ami => "ami-12345678"

    # More comprehensive region config
    aws.region_config "us-west-2" do |region|
      region.ami = "ami-87654321"
      region.keypair_name = "company-west"
    end
  end
end

The region-specific configurations will override the top-level configurations when that region is used. They otherwise inherit the top-level configurations, as you would probably expect.

Networking features in the form of config.vm.network are not supported with vagrant-aws, currently. If any of these are specified, Vagrant will emit a warning, but will otherwise boot the AWS machine.

There is minimal support for synced folders. Upon vagrant up, vagrant reload, and vagrant provision, the AWS provider will use rsync (if available) to uni-directionally sync the folder to the remote machine over SSH.

See Vagrant Synced folders: rsync

To use tags, simply define a hash of key/value for the tags you want to associate to your instance, like:

Vagrant.configure("2") do |config|
  # ... other stuff

  config.vm.provider "aws" do |aws|
    aws.tags = {
	  'Name' => 'Some Name',
	  'Some Key' => 'Some Value'
    }
  end
end

You can specify user data for the instance being booted.

Vagrant.configure("2") do |config|
  # ... other stuff

  config.vm.provider "aws" do |aws|
    # Option 1: a single string
    aws.user_data = "#!/bin/bash\necho 'got user data' > /tmp/user_data.log\necho"

    # Option 2: use a file
    aws.user_data = File.read("user_data.txt")
  end
end

Need more space on your instance disk? Increase the disk size.

Vagrant.configure("2") do |config|
  # ... other stuff

  config.vm.provider "aws" do |aws|
    aws.block_device_mapping = [{ 'DeviceName' => '/dev/sda1', 'Ebs.VolumeSize' => 50 }]
  end
end

You can automatically attach an instance to an ELB during boot and detach on destroy.

Vagrant.configure("2") do |config|
  # ... other stuff

  config.vm.provider "aws" do |aws|
    aws.elb = "production-web"
  end
end

To work on the vagrant-aws plugin, clone this repository out, and use Bundler to get the dependencies:

$ bundle

Once you have the dependencies, verify the unit tests pass with rake:

$ bundle exec rake

If those pass, you're ready to start developing the plugin. You can test the plugin without installing it into your Vagrant environment by just creating a Vagrantfile in the top level of this directory (it is gitignored) and add the following line to your Vagrantfile

Vagrant.require_plugin "vagrant-aws"

Use bundler to execute Vagrant:

$ bundle exec vagrant up --provider=aws