OK, I think I've catched it. I can reproduce the problem when I override the remote user in the playbook file:
- hosts: all user: *root* *...* In my experience, the precedence rules for the applied ssh remote user are following: 1. ansible_ssh_user in inventory file (or defined as extra variable<https://github.com/al3x/sovereign/blob/b77bcdbb0c7fa7e66d4e51e17ed88ede56614fee/Vagrantfile#L32> ) 2. *user YAML attribute in playbook file (recently discovered ;-)* 3. -u / --user ansible-playbook argument (used by Vagrant provisioner) 4. ANSIBLE_REMOTE_USER environment variable 5. remote_user paramter in an ansible.cfg file 6. username of the current user At the moment, Vagrant behavior can therefore be shadowed by configuration 1 and 2, which is the origin of this confusion. I'll re-open your pull request to evaluate, which strategy makes more sense... -- You received this message because you are subscribed to the Google Groups "Ansible Project" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
