Increase EC2 Resource Count in Terraform and applying Ansible

Question

Having some issue with increasing the EC2 resources for terraform and applying the changes with Ansible. I currently have a basic template with:

resource "aws_instance" "linux_node" {
  count = 2
  ami           = "ami-6df1e514"
  instance_type = "t2.micro"
  key_name = ""
  vpc_security_group_ids = [""]
  iam_instance_profile = ""
  user_data = "${file("userdata.sh")}"

  tags {
    Name = "linux-node"
  }

  provisioner "local-exec" {
        command = "sleep 120; ANSIBLE_HOST_KEY_CHECKING=False ansible-playbook --private-key ~/.ssh/aws-patrick-oregon.pem -i '${aws_instance.linux_node.public_ip},' ../install_packages.yml"
    }
}

The issue when I increase the count to anything more than 1, I get an error on aws_instance.linux_node.public_ip where the error says:

* aws_instance.linux_node[1]: 1 error(s) occurred:

* Resource 'aws_instance.linux_node' not found for variable 'aws_instance.linux_node.public_ip'

I think it stems from the bracket [1]. I'm wondering how to write a ansible command to pick up the change when I make the change to add an EC2 instance.

Answer 1

When a resource has count set, it needs to be addressed in a different way to specify which of the instances you wish to use.

For provisioners in particular it's a common case to want to take actions based on the "current" instance the provisioner is running for, so a shorthand is provided with the self variable type:

provisioner "local-exec" {
  command = "sleep 120; ANSIBLE_HOST_KEY_CHECKING=False ansible-playbook --private-key ~/.ssh/aws-patrick-oregon.pem -i '${self.public_ip},' ../install_packages.yml"
}