Caching variable name in .each

[mattchilds1]

Hi, I have a problem when I am creating multiple AWS machines. Locally they look like:

When I look on Amazon they unfortunately all have the name 'node1-aws'. Looking at the --debug output, it seems that it is using 'With machine: node1-aws' for every node.

 INFO machine: Initializing machine: node3-aws
...
INFO machine: New machine ID: nil
 INFO command: With machine: node1-aws (#<VagrantPlugins::AWS::Provider:0x00000102869bd0 @machine=#<Vagrant::Machine: node1-aws (VagrantPlugins::AWS::Provider)>

My vagrant file:

# -*- mode: ruby -*-
# vi: set ft=ruby :

boxes = [
    {
        :name => "node1-aws",
        :playbook => "remote-bootstrap.yml",
        :region => "eu-west-1",
        :ami => "ami-3910c44e",
        :instance_type => "t2.micro",
        :ansible_inventory => "amazon-ec2-inventory"
    },
    {
        :name => "node2-aws",
        :playbook => "remote-bootstrap.yml",
        :region => "eu-west-1",
        :ami => "ami-3910c44e",
        :instance_type => "t2.micro",
        :ansible_inventory => "amazon-ec2-inventory"
    },
    {
        :name => "node3-aws",
        :playbook => "remote-bootstrap.yml",
        :region => "eu-west-1",
        :ami => "ami-3910c44e",
        :instance_type => "t2.micro",
        :ansible_inventory => "amazon-ec2-inventory"
    }
]

Vagrant.configure(2) do |config|
  config.vm.box = "etailer/ubuntu-13.10"

  boxes.each do |opts|
    config.vm.define opts[:name] do |node|
      config.vm.provider :aws do |aws, override|
        override.ssh.private_key_path = "~/.ssh/id_rsa"
        aws.access_key_id = "removed"
        aws.secret_access_key = "removed"
        aws.region = opts[:region]
        aws.keypair_name = "id_rsa"
        aws.ami = "ami-3910c44e"
        aws.instance_type = opts[:instance_type]
        aws.tags = {
          'Name' => opts[:name]
        }
        #override.vm.box = "etailer/ubuntu-13.10"
      end #end of AWS

      node.vm.provision "ansible" do |ansible|
        ansible.verbose = "v"
        ansible.playbook = "../remote-bootstrap.yml"
        ansible.inventory_path = "../amazon-ec2-inventory"
        ansible.host_key_checking = "true"
        ansible.limit = "all"
      end # end of ansible for aws

    end #end of node

  end #end of loop

end #end of vagrantfile

Is there some way to stop this cacheing? or am I just doing something stupid? :dancers:

[hrpatel]

Did you ever figure this out? Seems I've just run into the same issue.

More specifically, I'm getting the same AMI and Tags for each instance:

...
# Read in the JSON configuration
test_boxes = JSON.parse(
  File.read(
    File.join(File.dirname(__FILE__), config_file)
  ),
  symbolize_names: true
)

...

# Start Vagrant configuration
Vagrant.configure(VAGRANTFILE_API_VERSION) do |vagrant|

  ...

  # Loop through each box we define in the JSON file
  test_boxes.each_pair do |vm_name, box_attributes|

      ...

      # Configure instance specific AWS variables
      vagrant.vm.provider :aws do |aws, override|
        aws.ami = box_attributes[:ami]
        aws.tags = {
          # FIXME: all instances get the name of the first instance
          'Name' => [ENV['USER'], vm_name].join('-'),
          'Vagrant' => true
        }
...