# ansible 2.1 setup fails with this module requires key=value arguments (\['\<\\>'\]) **URL:** \ **Category:** Ansible Project\ **Tags:** ubuntu\ **Created:** [June 5, 2016, 4:12pm UTC](https://forum.ansible.com/t/ansible-2-1-setup-fails-with-this-module-requires-key-value-arguments-include-ansible-module-args/21667 "2016-06-05T16:12:45Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![Greg\_Fausak](https://avatars.discourse-cdn.com/v4/letter/g/a88e4f/32.png) [@Greg\_Fausak](https://forum.ansible.com/u/Greg_Fausak)\ **Post date:** [June 5, 2016, 4:12pm UTC](https://forum.ansible.com/t/ansible-2-1-setup-fails-with-this-module-requires-key-value-arguments-include-ansible-module-args/21667/1 "2016-06-05T16:12:45Z")
I am migrating to 2.1 from 1.9. I ran into a few issues, but it has been fairly painless. I’ve run across this, and I have narrowed it down to exactly how to recreate it. Once I get in this state ansible is not usable anymore. I’ve got a playbook, target is ubuntu 15.10, source is mac osx. i am using vagrant to setup a vm. then i ssh in modifying the ansible.cfg: [default] nocows = yes retry\_files\_enabled = False remote\_user = vagrant deprecation\_warnings = False host\_key\_checking = False force\_handlers = True #allow\_world\_readable\_tmpfiles = True hostfile = /Users/c0faulo/git/vagrant/hosts [ssh\_connection] ssh\_args = -F /Users/c0faulo/git/vagrant/machines/blank/.vagrant\_ssh\_config -o ControlMaster=auto -o ControlPersist=30m scp\_if\_ssh = True control\_path = ~/.ssh/ansible-%%r@%%h:%%p If it matters, here is my .vagrant\_ssh\_config file: Host default HostName 127.0.0.1 User vagrant Port 2222 UserKnownHostsFile /dev/null StrictHostKeyChecking no PasswordAuthentication no IdentityFile “/Users/c0faulo/git/vagrant/machines/blank/.vagrant/machines/default/virtualbox/private\_key” IdentitiesOnly yes LogLevel FATAL Everything works in the playbook. When I try to run the playbook again I get this error. By binary search I found the line that messes up my ansible connection: I have several repos from my accounts that I install, then I run a command (usually python setup.py install) in that repo. Here is the task: - name: install repos become: yes become\_user: “{{item.user}}” shell: “( cd ~{{login}}/git/{{item.repo}}; {{item.install}})” with\_items: “{{ repos }}” and my repos dictionary: vars: repos: - repo: “gen\_yaml” org: “VCP” install: “echo nothing to do” #install: “python setup.py install” user: “root” url: “[http://txslnno-github.cds.eng.vzwcorp.com/VCP/gen\_yaml.git](http://txslnno-github.cds.eng.vzwcorp.com/VCP/gen_yaml.git)” - repo: “dhcp-pb” org: “VCP” install: “echo nothing to do” user: “{{login}}” url: “[http://txslnno-github.cds.eng.vzwcorp.com/VCP/dhcp-pb.git](http://txslnno-github.cds.eng.vzwcorp.com/VCP/dhcp-pb.git)” - repo: “cmdb” org: “corona” install: “make site\_torrance” user: “{{login}}” url: “[http://txslnno-github.cds.eng.vzwcorp.com/corona/cmdb.git](http://txslnno-github.cds.eng.vzwcorp.com/corona/cmdb.git)” you can see the install: “echo nothing to to” When I change that to the python setup.py install, everything works, but then I can’t run ansible again. Something appears to get messed up in the ansible environment on the target side. My program ‘gen\_yaml’ is a fairly simple program. It needs pyyaml and psycopg2. I did a manual install and this is the verbose output : creating dist creating ‘dist/gen\_yaml-0.1.9-py2.7.egg’ and adding ‘build/bdist.linux-x86\_64/egg’ to it removing ‘build/bdist.linux-x86\_64/egg’ (and everything under it) Processing gen\_yaml-0.1.9-py2.7.egg creating /usr/local/lib/python2.7/dist-packages/gen\_yaml-0.1.9-py2.7.egg Extracting gen\_yaml-0.1.9-py2.7.egg to /usr/local/lib/python2.7/dist-packages Adding gen-yaml 0.1.9 to easy-install.pth file Installing gen\_yaml script to /usr/local/bin Installed /usr/local/lib/python2.7/dist-packages/gen\_yaml-0.1.9-py2.7.egg Processing dependencies for gen-yaml==0.1.9 Searching for psycopg2==2.6 Best match: psycopg2 2.6 Adding psycopg2 2.6 to easy-install.pth file Using /usr/lib/python2.7/dist-packages Finished processing dependencies for gen-yaml==0.1.9 My program gen\_yaml works, and the installation worked, but I can’t access with the ansible anymore? -g ---
**Author:** ![Greg\_Fausak](https://avatars.discourse-cdn.com/v4/letter/g/a88e4f/32.png) [@Greg\_Fausak](https://forum.ansible.com/u/Greg_Fausak)\ **Post date:** [June 6, 2016, 7:13pm UTC](https://forum.ansible.com/t/ansible-2-1-setup-fails-with-this-module-requires-key-value-arguments-include-ansible-module-args/21667/2 "2016-06-06T19:13:39Z")
In the last 24 hours I looked all over for this. If I install _anything_ with python setup.py install I get a failure. The hint was the output in the setup.py process: Adding psycopg2 2.6 to easy-install.pth file /usr/local/lib/python2.7/dist-packages/easy-install.pth I realized I didn’t know what that meant, so I started looking around a little. It turns out there is a file called easy-install.pth that is read by python at a very low level. The contents of my file after Ansible stopped working (after my setup.py): import sys; sys.\_\_plen = len(sys.path) ./cmdb\_utils-0.0.4-py2.7.egg ./argparse\_config-0.5.1-py2.7.egg /usr/lib/python2.7/dist-packages import sys; new = sys.path[sys.\_\_plen:]; del sys.path[sys.\_\_plen:]; p = getattr(sys, ‘\_\_egginsert’, 0); sys.path[p:p] = new; sys.\_\_egginsert = p + len(new) And the contents of the file before I do a setup.py: If I edit the file and remove this line: /usr/lib/python2.7/dist-packages Then ansible works again. It has something to do with the order of my apt-gets and my python setup.py installs. II haven’t been able to figure out exactly how to recreate it. -g ---
**Author:** ![Roman\_Gaufman1](https://avatars.discourse-cdn.com/v4/letter/r/c6cbf5/32.png) [@Roman\_Gaufman1](https://forum.ansible.com/u/Roman_Gaufman1)\ **Post date:** [July 28, 2016, 11:21am UTC](https://forum.ansible.com/t/ansible-2-1-setup-fails-with-this-module-requires-key-value-arguments-include-ansible-module-args/21667/3 "2016-07-28T11:21:58Z")
Hi Greg, I’m experiencing the same thing on 2 servers, rest are fine. I can ssh to them and I cannot see anything obviously wrong: TASK [setup] \*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\* fatal: [0ca87760df7de0fe2d36]: FAILED! =\> {“changed”: false, “failed”: true, “msg”: “this module requires key=value arguments ([‘\<\\>’])”} fatal: [33ac8c46007794dee1a6]: FAILED! =\> {“changed”: false, “failed”: true, “msg”: “this module requires key=value arguments ([‘\<\\>’])”} to retry, use: --limit @timeboxes.retry I resolved the problem by moving this file out of the way: sudo mv /usr/local/lib/python2.7/dist-packages/easy-install.pth /usr/local/lib/python2.7/dist-packages/easy-install.pth.backup Very strange. ---
**Author:** ![Adam\_Grant](https://avatars.discourse-cdn.com/v4/letter/a/ee59a6/32.png) [@Adam\_Grant](https://forum.ansible.com/u/Adam_Grant)\ **Post date:** [August 1, 2016, 11:28pm UTC](https://forum.ansible.com/t/ansible-2-1-setup-fails-with-this-module-requires-key-value-arguments-include-ansible-module-args/21667/4 "2016-08-01T23:28:35Z")
Hi Roman, I believe you are seeing this: FAILED! =\> {“changed”: false, “failed”: true, “msg”: “this module requires key=value arguments ([‘\<\\>’])”} because those 2 servers are sourcing local out-of-date ansible.egg files (in their easy-install.pth)? It was the case for me. This seems to be creating a namespace collision for Ansible 2.1 lib imports when the remote python interpreter is initialized to run python code remotely. -Adam