Correct workflow for new ansible deploy

[snoby]

I've gone through the docs and reviewed the reasons and the work done to change to the new ansible deploy methodology. Great job, I'm sure that was a lot of work. My question is if I want to work out of the source code tree with ansible, I'm failing to see what is the proper way to run the playbooks. Pretend that i don't want to use ansible collections.

chatgpt is telling me to do some magic with paths and fake the playbooks like they are a collection. Or i could change all the roles that have - role: vitabaks.autobase.xxxx to just - role: xxx

Just wondering what your workflow is or how you think my workflow should be if i want to work on the playbooks or debug them.

thanks.

[vitabaks]

Hi!

We are currently transitioning to the Ansible Collection format, and this was requested by many users. We even received some feedback that directly using the source code tree is not considered a best practice in the Ansible ecosystem, and we’ve taken this into account.

Up to version 2.2.0 (inclusive), you can still use your usual workflow and run the playbooks directly from the source code as before.
However, starting with the next release, we plan to recommended only three approaches:

1. Using the Autobase Console (UI)
2. Using the autobase/automation Docker image, which already includes all the necessary code and dependencies
3. Installing and using as an Ansible Collection via ansible-galaxy

[vitabaks]

chatgpt is telling me to do some magic with paths and fake the playbooks like they are a collection. Or i could change all the roles that have - role: vitabaks.autobase.xxxx to just - role: xxx

Perhaps this will help: in the Docker image, we make the source code available as an Ansible Collection simply by creating a symlink.

[snoby]

Yes i think the symlink is exactly what i'm looking for. Great job with the deployment modifications!

I'm working on implementing the autobase deployment to move away from a docker based repmgr setup from bitnami. It's super fragile. We are moving out of AWS and into colocation so going to try and do cross region replicas as well... This is all on a system that has zero downtime allowed. So will be live migration ( we've done that 3 times already). Fingers crossed! ( Of course will be doing lots of lab testing first).

[pat-s]

@snoby Assuming you have installed the collection from Ansible Galaxy:

- name: Autobase
  hosts: etcd_cluster
  become: true
  gather_facts: true
  any_errors_fatal: true
  become_user: root


- name: Run Autobase config_pgcluster
  ansible.builtin.import_playbook: vitabaks.autobase.config_pgcluster

See also the README section for this.