Closed msimonin closed 6 years ago
Any reasons why this is done asynchronously? I mean, another solution would be to add detach: False
in the docker_container
rule to execute it synchronously.
(Sorry, I am a bit late on that one)
Oh yes detach: False
would have done the trick.
Should we change it ?
I vote yes since the task will reflect the status of the container. Thus, enos init
will fail if the init_os
fails. Could not remember if this also avoids docker logs and debug tasks from https://github.com/BeyondTheClouds/enos/commit/3ecd3988b2df4ed2b905f4c0a060473569d720eb
Agreed, I'd keep the debug if that's possible.
Currently init is launched asynchronously. This can lead to race when the deployment is used directly without waiting for all the initial resources to be created.
see: https://github.com/BeyondTheClouds/enos/blob/8510f44bb418f4774ddfc5a393a7c04314ac0d93/enos/ansible/roles/init_os/tasks/main.yml#L39-L46