Starting Stopped AWS EC2 Instances from Ansible Tower

[John Peters]

Any clue as to how to accomplish this?

My Ansible Tower inventory has a list of hosts in it, all in 'stopped' state.

I'd like to use that inventory, assign it to a playbook, that will loop through each machine in the inventory, from the localhost, and issue a command to start the instance to AWS.

I've tried several different ways to do this, but to no avail.

It seems that because I'm specifying an inventory to the job, and that inventory only has stopped instances in it, there's no way to execute a task on the localhost, loop through that inventory, and start them, one by one?

Here is some code that contains all the different lines I've tried...

---

- name: Start EC2 Instances

  hosts: localhost # also tried 'all' here

  connection: local

  gather_facts: false

  tasks:

    - name: inventory_hostname var

      debug:

        msg: "{{ hostvars[inventory_hostname] }}"

      with_inventory_hostnames: all

      delegate_to: localhost

    - name: hostvars item

      debug:

        msg: "{{ hostvars[item] }}"

      with_inventory_hostnames: all

      delegate_to: localhost

   - name: Start EC2 Instances Task

     ec2_instance: running

      local_action:

        module: ec2

          instance_ids={{hostvars[inventory_hostname]['ec2_id']}}

          region="{{ region | default('us-east-1') }}"

          state=running

     when: hostvars[item]['ec2_id'] is defined

     with_inventory_hostnames: all

     delegate_to: localhost

[Kuldip Madnani]

If gather_facts is false and node is unreachable, how are you going to retrieve hostvars[item]['ec2_id'] ? Is it statically declared in your inventory file?