When wanting to make changes to ansible.cfg do I make them on the web or task container? Also is there any documentation as to what should be done where and when? For anything the Ansible tower doc’s say to edit/change is that all on the web or task container? Some thing’s I’ve made changes too don’t seem to be applied? Do I need to cycle containers for any config changes?
Thanks
It depends, my understanding is Ansible template/workflow executions are fully run on the task containers whereas UI stuff is executed on the web container, the caveat being that some of the UI stuff uses Ansible (such as getting the version of Ansible). My recommendation is change the config in both places to be safe and ensure consistency. Although if you’re looking for user level customizations, it might be easier to ask users to supply their own config in their projects instead of changing the system level config.
As for the need to cycle containers – I’m not sure. Keep in mind that if you do cycle the containers, changes to ansible.cfg will not persist unless that change exists in the image itself.
So if I needed to cycle the container for any reason should changes also be made to the rhel/server hosting the docker container. On my server I see /etc/ansible/ansible.cfg along with a bunch of docker paths with asnible.cfg files.
I’m assuming from your statement that during a cycle it reads from the source pulled from github not the files on the container?
Also when you makes changes in the UI IE: the ISOLATED_COLLETION_TIMEOUT value from 10 to 5. Does that trickle down to some file on one of the containers or is that stored/updated in the DB.
Thanks for your help.
Well, assuming you’re using containerized AWX, configuration changes you make on the host will not affect any processes running in the containers. If you want to change the cfg on a container, you’d need to log into the container and make the changes on the container’s filesystem.
Cycle behavior depends on your use case - if you’re using Swarm, doing a docker restart will simply restart the containers. But if you do a docker-compose down && docker-compose up -d, the containers will look for the web/task images (either locally or on DockerHub) and use those when starting up, effectively resetting any changes you’d made to the running containers. K8s does this differently too.
And I believe changes made in the UI are persisted in the DB as containers are ephemeral.