AWS AMI Cleanup w/Ansible iterate through results array - amazon-web-services

Question

I have a previous task that creates weekly backups, labeling them with the server name followed by a date/time tag. The goal of this job is to go in behind it and clean up the old AMI backups, leaving only the last 3. The ec2_ami_find task works fine, but it could also return empty results for some servers and I'd like the deregister task to handle that.
The error I'm getting is pretty generic:
fatal: [127.0.0.1]: FAILED! => {
"failed": true,
"msg": "The conditional check 'item.ec2_ami_find.exists' failed. The error was: error while evaluating conditional
(item.ec2_ami_find.exists): 'dict object' has no attribute
'ec2_ami_find'\n\nThe error appears to have been in
'/root/ansible/ec2-backups-purge/roles/first_acct/tasks/main.yml': line 25,
column 3, but may\nbe elsewhere in the file depending on the exact
syntax problem.\n\nThe offending line appears to be:\n\n\n- name:
Deregister old backups\n ^ here\n"
The playbook task reads as follows:
---
- name: Find old backups
tags: always
ec2_ami_find:
owner: self
aws_access_key: "{{ access_key }}"
aws_secret_key: "{{ secret_key }}"
region: "{{ aws_region }}"
ami_tags:
Name: "{{ item }}-weekly-*"
sort: name
sort_order: descending
sort_start: 3
with_items:
- server-01
- server-02
- server-win-01
- downloads
register: stale_amis
- name: Deregister old backups
tags: always
ec2_ami:
aws_access_key: "{{ access_key }}"
aws_secret_key: "{{ secret_key }}"
region: "{{ aws_region }}"
image_id: "{{ item.ami_id }}"
delete_snapshot: True
state: absent
with_items:
- "{{ stale_amis.results }}"
Snippet of one of the results returns:
"results": [
{
"ami_id": "ami-zzzzzzz",
"architecture": "x86_64",
"block_device_mapping": {
"/dev/xvda": {
"delete_on_termination": true,
"encrypted": false,
"size": 200,
"snapshot_id": "snap-xxxxxxxxxxxxx",
"volume_type": "gp2"
}
},
"creationDate": "2017-08-01T15:26:11.000Z",
"description": "Weekly backup via Ansible",
"hypervisor": "xen",
"is_public": false,
"location": "111111111111/server-01.example.com-20170801152611Z",
"name": "server-01.example.com-20170801152611Z",
"owner_id": "111111111111",
"platform": null,
"root_device_name": "/dev/xvda",
"root_device_type": "ebs",
"state": "available",
"tags": {
"Name": "server-01-weekly-20170801152611Z",
"Type": "weekly"
},
"virtualization_type": "hvm"
},

Answer 1

I doubt your attempt:
with_items:
- "{{ stale_amis.results }}"
because ec2_ami_find put results into own results field. So the first AMI for first server will be stale_amis.results[0].results[0].ami_id
I advice to reduce original stale_amis to required list and loop over it. For example you can use json_query filter:
- ec2_ami:
aws_access_key: "{{ access_key }}"
aws_secret_key: "{{ secret_key }}"
region: "{{ aws_region }}"
image_id: "{{ item }}"
delete_snapshot: True
state: absent
with_items: "{{ stale_amis | json_query('results[].results[].ami_id') }}"