I recently got some errors on the end nodes I was deploying to. Error is: “-su: /dev/null: Permission denied”, which in general seems to be recoverable after a reboot. It looks like the /dev/null file is just deleted somehow. Has anyone else seen this? Very weird.
Can you please share the relevant command lines and playbook snippets that got you there? (Also can we see the output in the context of the run?)
Ansible version?
I see su, so I’m not sure if you are using the new 1.5 su keyword or are running su instead the playbook, etc.
Thanks!
Can you please share the relevant command lines and playbook snippets that
got you there?
It's a pretty big role-based system. Can I share it with you privately?
(Also can we see the output in the context of the run?)
I don't think I have it, unfortunately. It happened several times, but not
always. Never had this problem before with the same system. I kind of
wonder if it correlates with me CTRL-C exiting the ansible-playbook run...
Ansible version?
Not sure exactly. I have recently upgraded to 1.4.4 via pip. But I was
running maybe 1.2.3 before that (like a day ago). Unfortunately I am not
sure if I got the problem on one or the other or both.
I can try deploying to a local node tonight to see if I can recreate the
problem.
I see su, so I'm not sure if you are using the new 1.5 su keyword or are
running su instead the playbook, etc.
I don't think I use su anywhere explicitly (shelled out), but my roles do
sometimes change user.
This turned out to be a script error, where I was rm’ing a file that occassionally is /dev/null Ansible was in no way responsible. Sorry for the false alarm and thanks for the awesome software!