redfish_facts results an Unknown error

[warusadura]

SUMMARY

Hello Ansible community!
When executing a simple playbook with redfish_facts module, results Unknown error

fatal: [HOST]: FAILED! => {
    "changed": false,
    "invocation": {
        "module_args": {
            "baseuri": "http://127.0.0.1:8000/",
            "category": [
                "Manager"
            ],
            "command": [
                "all"
            ],
            "password": "VALUE_SPECIFIED_IN_NO_LOG_PARAMETER",
            "username": "foo"
        }
    },
    "msg": "Unknown error"
}

I'm unable to debug/improve my playbook with an error like Unknown error. So, I think it would be great if we could improve this to much more verbose error message.

Thanks :)

ISSUE TYPE

• Bug Report

COMPONENT NAME

redfish_facts

ANSIBLE VERSION

ansible 2.7.4
python version = 3.7.1 (default, Oct 21 2018, 08:03:53) [GCC 8.2.0]

[ansibot]

Files identified in the description:

• lib/ansible/modules/remote_management/redfish/redfish_facts.py

If these files are inaccurate, please update the component name section of the description or use the !component bot command.

click here for bot help

[ansibot]

cc @billdodd @jose-delarosa @mraineri @tomasg2012
click here for bot help

[jose-delarosa]

@dnuka I agree that some of the error messages could be more specific. For now, can you please provide more information about your environment? What out-of-band device is listening at http://127.0.0.1:8000? Did you modify the redfish_facts module so that it would connect using http instead of https? (default is https).

[warusadura]

Hi @jose-delarosa. Thank you for the quick response. I'm an intern at the OpenStack Ironic team.

For now, can you please provide more information about your environment? What out-of-band device is listening at http://127.0.0.1:8000?

I have configured to run sushy-emulator on a remote cent-os server. And ran this playbook against it.

Did you modify the redfish_facts module so that it would connect using http instead of https? (default is https).

No. I didn't modify the redfish_facts module.

I'm very much new to IT automation, Redfish and still learning Ansible :) Right now I'm trying to figure out why Ansible is outputting "Unknown error". Also, I'm really interested in contributing to redfish_facts module with generating a more specific error message.

Thanks.

[billdodd]

Hi, @dnuka.

Thanks for opening the issue. The "Unknown error" is coming from the redfish_utils module in the final except block of one of the methods get_request(), post_request(), patch_request(), or delete_request(). A good improvement here would be to output a representation of the Exception received rather than just 'Unknown error'. If you are interested, feel free to submit a PR for a fix like that.

[warusadura]

Hi :) @billdodd Thank you for the response.

I was able to figure out why Ansible is outputting "Unknown error" on my side. I have made a mistake when passing baseuri. Also, I just updated the redfish_facts module to support http @jose-delarosa

Updated playbook
Now I'm getting the following error message,

fatal: [HOST]: FAILED! => {
    "changed": false,
    "module_stderr": "Shared connection to HOST closed.\r\n",
    "module_stdout": "Traceback (most recent call last):\r\n  File \"/root/.ansible/tmp/ansible-tmp-1544508444.8116581-100042417563386/AnsiballZ_redfish_facts.py\", line 113, in <module>\r\n    _ansiballz_main()\r\n  File \"/root/.ansible/tmp/ansible-tmp-1544508444.8116581-100042417563386/AnsiballZ_redfish_facts.py\", line 105, in _ansiballz_main\r\n    invoke_module(zipped_mod, temp_path, ANSIBALLZ_PARAMS)\r\n  File \"/root/.ansible/tmp/ansible-tmp-1544508444.8116581-100042417563386/AnsiballZ_redfish_facts.py\", line 48, in invoke_module\r\n    imp.load_module('__main__', mod, module, MOD_DESC)\r\n  File \"/tmp/ansible_redfish_facts_payload_sIgx0e/__main__.py\", line 268, in <module>\r\n  File \"/tmp/ansible_redfish_facts_payload_sIgx0e/__main__.py\", line 199, in main\r\n  File \"/tmp/ansible_redfish_facts_payload_sIgx0e/ansible_redfish_facts_payload.zip/ansible/module_utils/redfish_utils.py\", line 133, in _find_systems_resource\r\nUnboundLocalError: local variable 'systems_service' referenced before assignment\r\n",
    "msg": "MODULE FAILURE\nSee stdout/stderr for the exact error",
    "rc": 1
}

Thanks for opening the issue. The "Unknown error" is coming from the redfish_utils module in the final except block of one of the methods get_request(), post_request(), patch_request(), or delete_request()

@billdodd Can you please provide me the link to redfish_utils module :)

A good improvement here would be to output a representation of the Exception received rather than just 'Unknown error'. If you are interested, feel free to submit a PR for a fix like that.

Yes! I'm extremely interested.

Thanks.

[billdodd]

The redfish_facts module is here:

https://github.com/ansible/ansible/blob/devel/lib/ansible/module_utils/redfish_utils.py

For the new error you are seeing, it looks like the Members array in your Systems resource is either empty or missing.

[dtantsur]

I think both errors are bugs. Even if the server is not fully compliant (which I'm not sure), the errors should be better.

[warusadura]

Thank you @billdodd :)

For the new error you are seeing, it looks like the Members array in your Systems resource is either empty or missing.

It appears so,

{
    "@odata.type": "#ComputerSystemCollection.ComputerSystemCollection",
    "Name": "Computer System Collection",
    "Members@odata.count": 0,
    "Members": [
        
    ],
    "@odata.context": "/redfish/v1/$metadata#ComputerSystemCollection.ComputerSystemCollection",
    "@odata.id": "/redfish/v1/Systems",
    "@Redfish.Copyright": "Copyright 2014-2016 Distributed Management Task Force, Inc. (DMTF). For the full DMTF copyright policy, see http://www.dmtf.org/about/policies/copyright."
}

I agree with @dtantsur :) We could do better by providing more specific error messages to the end user.

Also, why there are two locations? I mean, lib/ansible/module_utils/ and then /lib/ansible/modules/ I thought only things under lib/ansible/modules were considered valid custom modules. Is this a design choice or is there a specific reason behind it :)

Thanks.

[billdodd]

I completely agree that it is a bug causing the UnboundLocalError when there are no systems. There is another issue related to multiple Systems (#47108). When I fix that issue, the UnboundLocalError will also be fixed.

Regarding the module_utils path, it's part of the Ansible architecture. There is a brief description here: https://docs.ansible.com/ansible/2.7/dev_guide/developing_module_utilities.html

Having much of the core Redfish function in a utility module allows vendors to extend that utility module if they need to develop a module that provides non-standard (OEM) function.

[warusadura]

Hi @billdodd

How do I test a modified module. I have found this. But there is no mention of testing against a remote server. Also, I couldn't find any tests for already existing custom redfish modules. Do you have some tests somewhere in some form? As a contributor how do I ensure the workings of the change?

Thanks :)

[billdodd]

Hi @dnuka - The link you referenced describes the environment setup I use for development and testing (at the beginning of the page in the "Environment setup" section). With that venv and hacking/env-setup, when you run a playbook it will use the modules in your local repository.

There aren't any tests (like unit tests or functional tests) for the Redfish ansible modules as far as I know.

But the sample inventory file and playbooks in this repo should be helpful for seeing how to test against a remote Redfish server: https://github.com/dell/redfish-ansible-module

[warusadura]

@billdodd Thank you very much for answering all my questions :) It was very helpful. I had some doubts about executing $ . hacking/env-setup. Your answer clears everything.

Please let me know, If there is anything Ansible and Redfish related. I'm very much happy to learn and contribute.
Thanks again :)

[gundalow]

Proposed fix in #50129 (thanks @dnuka) would welcome review on that