Unable to use the command module consecutively.

[megoix]

Ansible version: 1.0

Problem:

Using "action: command /usr/sbin/setenforce 0" in a playbook to disable SELinux only works once. Any consecutive runs after that will fail with the following message.

CentOS 6.3:
"failed: [el6_ansible_1] => {"changed": true, "cmd": ["/usr/sbin/setenforce", "0"], "delta": "0:00:00.001085", "end": "2013-02-12 10:36:39.363762", "rc": 1, "start": "2013-02-12 10:36:39.362677"}
stderr: /usr/sbin/setenforce: SELinux is disabled"

CentOS 5.9:
"failed: [el5_ansible_1] => {"changed": true, "cmd": ["/usr/sbin/setenforce", "0"], "delta": "0:00:00.000992", "end": "2013-02-12 10:36:39.294839", "rc": 1, "start": "2013-02-12 10:36:39.293847"}
stderr: /usr/sbin/setenforce: SELinux is disabled"

Expected behavior:

That it will behave like it does from the terminal, set SELinux to disabled every time it is run. Which it in a sense does, but from an ansible perspective it fails.

Workaround:

Install the python dependency for selinux and use the selinux module.

Test bed:

CentOS 5.9 and CentOS 6.3

[mpdehaan]

Use ignore_errors: True in this case or adapt your command string to only run if the command is in the right state.

It's not a fault of ansible that the command returned a non-zero return code.