Save playbook information to hosts file - ansible

Question

I have a playbook that spins up a new droplet on DigitalOcean using the core module built into Ansible:
- name: Provision droplet on DigitalOcean
local_action: digital_ocean
state=present
ssh_key_ids=1234
name=mydroplet
client_id=ABC
api_key=ABC
size_id=1
region_id=2
image_id=3
wait_timeout=500
register: my_droplet
- debug: msg="ID is {{ my_droplet.droplet.id }}"
- debug: msg="Your droplet has the IP address of {{ my_droplet.droplet.ip_address }}"
I run this using (note the local argument):
ansible-playbook playbooks/create_droplet.yml -c local -i playbooks/hosts
My hosts file initially looks like this:
[production]
TBA
[localhost]
localhost
When the above playbook finishes I can see the debug information in STDOUT:
ok: [localhost] => {
"msg": "Your droplet has the IP address of 255.255.255.255"
}
Is there any way for this playbook to retain the my_droplet.ip_address variable and save the TBA in the hosts file instead of having to manually copy-pasta it there? I ask because I want to add this provisioning playbook to a ruby script that subsequently bootstraps the VPS with another playbook.

Answer 1

I am doing the same, having written a play that creates servers from a dict (approximately 53 servers at a go, dynamically creating a full test environment). To use a static hosts file, I add the following:
- name: Create in-memory inventory
add_host:
name: "{{ item.value.ServerName }}"
groups: "{{ item.value.RoleTag }},tag_Environment_{{ env_name }}"
when: item.value.template is defined
with_dict: server_details
- name: create file
become: yes
template:
src: ansible-hosts.j2
dest: "{{ wm_inventory_file }}"
The ansible-hosts.j2 template is simply:
{% for item in groups %}
[{{item}}]
{% for entry in groups[item] %}
{{ entry }}
{% endfor %}
{% endfor %}

Answer 2

I'm launching instances with ec2, and I originally wanted to do the same thing. I found some examples of using lineinfile and massaging it to the do the same thing as so:
- name: Launching new instances for {{ application }}
local_action: ec2 group={{ security_group }} instance_type={{ instance_type}} image={{ image }} wait=true region={{ region }} keypair={{ keypair }} vpc_subnet_id={{ subnet }} count={{ instance_count }}
register: ec2
- name: Add instance to local host group
local_action: lineinfile dest=ansible_hosts regexp="{{ item.public_ip }}" insertafter="\[{{ application }}\]" line="{{ item.public_ip }} ansible_ssh_private_key_file=~/ec2-keys/{{ keypair }}.pem" state=present
with_items: ec2.instances
But, I have to agree that it's generally not something you want to do. I found it to be quite a pain. I've since switched to using add_host and life is much better. BTW, application would be the value I used for the group name...

Answer 3

Is there any way for this playbook to retain the my_droplet.ip_address
variable and save the TBA in the hosts file instead of having to
manually copy-pasta it there?
You can retain the ip address of your new host by using the add_host module which allows you to dynamically change the in-memory inventory during an ansible-playbook run. This is useful for when you want to provision a new host and then configure it in a single playbook.
For example
local_action: >
add_host
hostname={{ my_droplet.droplet.id }}
groupname=launched
And then later in your playbook:
- name: Configure instance(s)
hosts: launched
tasks:
...
the second part of your questions:
... and save the TBA in the hosts file instead of having to
manually copy-pasta it there?
There is no built-in ansible way to write additions to the inventory file that is on disk. This is generally not something you want to do. In this case you would need to add it or use a dynamic inventory script to discover the host for future configuration runs.

Answer 4

You should use a dynamic inventory script for this. Using name to distinguish instances, you can subsequently refer to your droplets.
Check https://github.com/ansible/ansible/blob/devel/contrib/inventory/digital_ocean.py for an example script.