Add support for passing of command line arguments (in a manner consistent with Ansible core modules)

[ktbyers]

---
- name: NAPALM on IOS 
  hosts: pynet-rtr1
  tasks:
    - name: NAPALM facts
      napalm_get_facts: 
        hostname: "{{ ansible_host }}"
        dev_os: "ios"

$ ansible-playbook napalm_w_cli.yml -u pyclass -k
SSH password: 

PLAY [NAPALM on IOS] *************************************************************************************************************************************************************************************************

TASK [Gathering Facts] ***********************************************************************************************************************************************************************************************
ok: [pynet-rtr1]

TASK [NAPALM facts] **************************************************************************************************************************************************************************************************
fatal: [pynet-rtr1]: FAILED! => {"changed": false, "failed": true, "msg": "username is required"}

PLAY RECAP ***********************************************************************************************************************************************************************************************************
pynet-rtr1                 : ok=1    changed=0    unreachable=0    failed=1   

Also should support if using a provider.

[varnumd]

Do you know if this feature was ever introduced - or do you know a good work-around?

[ktbyers]

I think that is done.

Let me know if it doesn't work though.

[varnumd]

Looks like I have the same error:

(ansible) [14:54:16] dvarnum:ansible-neteng git:(master*) $ ansible-playbook napalm_eos_testing.yml -l lab-campus-leaf3 -u dv -k                      
SSH password: 

PLAY [Get facts] *****************************************************************************************************************************************************************

TASK [get facts from device] *****************************************************************************************************************************************************
fatal: [lab-campus-leaf3]: FAILED! => {"changed": false, "msg": "username is required"}

PLAY RECAP ***********************************************************************************************************************************************************************
lab-campus-leaf3           : ok=0    changed=0    unreachable=0    failed=1    skipped=0    rescued=0    ignored=0   

Here's the playbook:

(ansible) [14:59:14] dvarnum:ansible-neteng git:(master*) $ cat napalm_eos_testing.yml
---
- name: Get facts
  hosts: eos
  connection: local
  gather_facts: no

  tasks:
    - name: get facts from device
      napalm_get_facts:
        hostname={{ inventory_hostname }}
        dev_os={{ ansible_network_os }}
        filter='facts,interfaces,bgp_neighbors'
      register: result    

    - name: print data
      debug: var=result

Am I missing something? Thanks for your time!

[ktbyers]

Hmmmm, I have to look into the username passing. I could have swore that worked, but I am seeing the same thing as you when I test it.

The -k passing works, but the -u username part is failing.

[ktbyers]

This fixes it...not sure why or when this changed though:

$ git diff ./napalm.py
diff --git a/napalm_ansible/plugins/action/napalm.py b/napalm_ansible/plugins/action/napalm.py
index ac37522..bbd70e0 100644
--- a/napalm_ansible/plugins/action/napalm.py
+++ b/napalm_ansible/plugins/action/napalm.py
@@ -13,7 +13,7 @@ class ActionModule(_ActionModule):
             provider = self._task.args.get('provider', {})
 
             provider['hostname'] = provider.get('hostname', provider.get('host', pc.remote_addr))
-            provider['username'] = provider.get('username', pc.connection_user)
+            provider['username'] = provider.get('username', pc.remote_user)
             provider['password'] = provider.get('password', pc.password)
             # Timeout can't be passed via command-line as Ansible defaults to a 10 second timeout
             provider['timeout'] = provider.get('timeout', 60)