When I run this simple Ansible playbook:
- name: EC2 Test Example
hosts: localhost
connection: local
gather_facts: False
tasks:
- name: EC2 Instance
ec2:
# Amazon EC2 key pair name
key_name: my-key-pair
# Amazon EC2 Security Group
group: my-security-group
instance_type: t2.micro
# Latest from https://wiki.debian.org/Cloud/AmazonEC2Image/Jessie
image: ami-221ea342
wait: yes
register: ec2
I run with venv/bin/ansible-playbook -i localhost, playbook.yml
:
PLAY [EC2 Test Example] ********************************************************
TASK [EC2 Instance] ************************************************************
fatal: [localhost]: FAILED! => {"changed": false, "failed": true, "msg": "boto required for this module"}
to retry, use: --limit @/Users/admin/temp/ansec2/playbook.retry
PLAY RECAP *********************************************************************
localhost : ok=0 changed=0 unreachable=0 failed=1
So obviously, I have boto installed in the venv that I'm using as well as my default system Python:
➜ ansec2 venv/bin/pip list
Package Version
--------------- --------
ansible 2.2.1.0
boto 2.45.0
boto3 1.4.4
botocore 1.5.4
...
I've read a few similar posts and I don't see a working solution.
The root cause of your problem is the -i localhost,
hack. You don't need to use it anymore in Ansible.
You can just run:
ansible-playbook playbook.yml
And with connection: local
in the play Ansible will use the Python executable set by venv.
When you use the -i localhost,
hack, Ansible calls its default /usr/bin/python
.
In this case you still can add the ansible_python_interpreter
parameter to tell Ansible to use this specific environment:
ansible-playbook -i localhost, playbook.yml --extra-vars "ansible_python_interpreter=/Users/admin/temp/ansec2/venv/bin/python"
But I think you should avoid it and use the first method.