How to do it...

We can create an EC2 instance using an Ansible module, which will take an instance type, AMI, subnet ID of VPC, tags, the exact count of instances, and region as input. Let's define tasks:

1. To log into the created instance, we need to create an EC2 keypair. There are two ways to create a keypair. One is that we create a keypair on the AWS console and download a private key; the other is to import an existing keypair into AWS. We will import a keypair into EC2:

- name: Create EC2 Key Pair
  ec2_key:
    name: my_first_key
    aws_access_key: "{{ access_key }}"
    aws_secret_key: "{{ secret_key }}"
    key_material: "{{ key }}"
    region: "{{ aws_region }}"
    state: present

Every AWS EC2 region maintains its own set of keypairs for instances to be launched in respective regions. We have passed aws_region, where we will be booting our instance and another variable that is key, which contains the content of our public key. This variable will be defined under ec2/vars/main.yml. After this task, our EC2 keypair will be generated and we can use the name defined for this keypair as the input in our next task to create an EC2 instance.    

1. We need to add the following things to our variables:

#Key Pair Content
key: 'ssh-rsa AAAAB3NzaC1yc2EAAAAD……………………….'

1. After we have our keypair in place, we are ready to create our EC2 instances in the subnets we created in VPC tasks. We have passed the following inputs to define our instance:

• • Instance configuration using the variable instance_type
  • Private subnet of the VPC we created using the registered variable; that is my_private.subnet.id in VPC task
  • Instance tags to refer our instance as key, value pair that is Name: Private Instance
  • Count of instances we want to keep in AWS EC2 region with specific tags using variable private_instance_count

The argument count_tag with exact_count determine how many nodes of an instance with the specific tag should be running. If the count of instances exceeds the parameters exact_count then extra instances would be terminated. In this example, we are keeping exact count for the instance tagged with Name: Private Instance.

- name: Create EC2 Instance in private subnet
  ec2:
    key_name: my_first_key
    instance_type: "{{ instance_type }}"
    image: "{{ ami_id }}"
    wait: yes
    group: my_first_sg
    vpc_subnet_id: "{{ my_private_subnet.subnet.id }}"
    aws_access_key: "{{ access_key }}"
    aws_secret_key: "{{ secret_key }}"
    region: "{{ aws_region }}"
    count_tag:
      Name: Private Instance
    exact_count: "{{ private_instance_count }}"
    instance_tags:
      Name: Private Instance

1. We need to define the following variables for the preceding tasks:

#EC2 Instance Info

instance_type: 't2.micro'
ami_id: 'ami-46c1b650'
private_instance_count: 1
public_instance_count: 1

1. Similarly, we can create an instance in the public subnet as well:

- name: Create EC2 Instance in public subnet
  ec2:
    key_name: my_first_key
    instance_type: "{{ instance_type }}"
    image: "{{ ami_id }}"
    wait: yes
    group: my_first_sg
    vpc_subnet_id: "{{ my_public_subnet.subnet.id }}"
    assign_public_ip: yes
    aws_access_key: "{{ access_key }}"
    aws_secret_key: "{{ secret_key }}"
    region: "{{ aws_region }}"
    count_tag:
      Name: Public Instance 
    exact_count: "{{ public_instance_count }}"
    instance_tags:
      Name: Public Instance
  register: ec2_public_instance

This task is identical to the previous task. The only difference is the different tags and subnet ID. We have created two instances in the preceding tasks, in each of the subnets we created in VPC tasks.