I am utterly confused about how to manage multiple environments(prod/dev) with ansible. I am also using molecule to test locally.
So here is my Project layout as of now.
|----inventories/
| |
| |--dev/
| | |
| | |--group_vars/...
| | |
| | |--host_vars/...
| |
| |--prod/
| |
| |--group_vars/...
| |
| |--host_vars/
| |
| |--my_playbook_hostname_vars.yml
|
|----roles/...
|
|----hosts.yml
|
|----my_playbook.yml
|
This is the directory structure according to the documentation
Now I have my molecule file where I can just link the dev directories like this
# molecule.yml
provisioner:
name: ansible
inventory:
links:
group_vars: ../../../../inventories/dev/group_vars/
host_vars: ../../../../inventories/dev/host_vars/
So Molecule works just fine for local testing with a vagrant driver and I could also set it up to use a EC2 driver to test it in the cloud for example. So far so good.
But how do I start the my_playbook.yml? When I use ansible-playbook my_playbook.yml then it does not know where to look for the vars, since there are two environments. How can I tell Ansible to look under inventories/prod/group_vars & inventories/prod/host_vars and then resolve further via the host and group name like it is defined in the hosts.yml without compromising my molecule setup?
When I just setup group_vars/ & host_vars/ in the root dir it works
According to documentation:
-i, --inventory
specify inventory host path or comma separated host list.
One option is to create a separate hosts file for each environment (pretty much the same way as in the documentation link you provided). Like this:
|-- inventories
| |-- dev
| | |-- group_vars/
| | |-- host_vars/
| | |-- hosts
| |-- prod
| |-- group_vars/
| |-- host_vars/
| |-- hosts
|-- roles/
|-- my_playbook.yml
...and then call ansible-playbook with -i
ansible-playbook my_playbook.yml -i inventories/dev/hosts
# or
ansible-playbook my_playbook.yml -i inventories/prod/hosts