Fix issue of the vmware_cluster_info module does not decode of a clus…

[sky-joker]

Depends-On: #374

SUMMARY

This PR is to fix the vmware_cluster_info module does not urldecode of a cluster name.
fixes: #365

ISSUE TYPE

• Bugfix Pull Request

COMPONENT NAME

plugins/modules/vmware_cluster_info

ADDITIONAL INFORMATION

tested on vCenter/ESXi 6.7

[RudneiBertolJr]

Hey @sky-joker,

Thank you for fix it.

I have a question, if it may be impacting other ansible resources that collect those facts from VMware, like Datacenter, Datastore, VM Guest name.

regards
rbertol

[sky-joker]

Hi @RudneiBertolJr

Thank you for reporting this issue.
When creating objects haven't this issue.
This issue occurs only when gathering information.
So some _info modules have the same issue.
I think the other _info modules need to be fixed in the same way.

First of all, can you please try this bug fixed module work at your environment, and please let me know the result?

[RudneiBertolJr]

Hey @sky-joker,

I have not tested yet, but I can tell you that it will work because it is very similar to the change that I did to work in my lab.

[root@dhcp181-132 ~]# cat /tmp/vmware_cluster_info.patch 
diff --git a/cloud/vmware/vmware_cluster_info.py b/cloud/vmware/vmware_cluster_info.py
index c1cf399..b3b1021 100644
--- a/cloud/vmware/vmware_cluster_info.py
+++ b/cloud/vmware/vmware_cluster_info.py
@@ -121,6 +121,11 @@ clusters:
 """
 
 try:
+    from urllib import unquote as urldecode
+except:
+    from urllib.parse import unquote as urldecode
+
+try:
     from pyVmomi import vim
 except ImportError:
     pass
@@ -213,7 +218,7 @@ class VmwreClusterInfoManager(PyVmomi):
                 vmware_client = VmwareRestClient(self.module)
                 tag_info = vmware_client.get_tags_for_cluster(cluster_mid=cluster._moId)
 
-            results['clusters'][cluster.name] = dict(
+            results['clusters'][urldecode(cluster.name)] = dict(
                 enable_ha=das_config.enabled,
                 ha_failover_level=ha_failover_level,
                 ha_vm_monitoring=das_config.vmMonitoring,

I will test it on Monday, and let you know if everything worked as expected.

regards
rbertol

[ansibullbot]

cc @Akasurde @Tomorrow9 @digifuchsi @goneri @lparkes @pgbidkar @warthog9
click here for bot help

[sky-joker]

By the way, you can even get around this issue with the following like playbook.
e.g., used urlencode filter

- name: Workaround Playbook Example
  hosts: localhost
  gather_facts: no
  vars:
    # urlencode will not convert from ‘/‘ to ‘%2f'
    # workaround: https://jinja.palletsprojects.com/en/2.11.x/templates/#urlencode
    cluster_name: "Cluster{{ '/\\%' | urlencode | replace('/', '%2f') | lower }}"
  tasks:
    - name: "Gather all cluster info"
      vmware_cluster_info:
        hostname: "{{ vcenter_hostname }}"
        username: "{{ vcenter_username }}"
        password: "{{ vcenter_password }}"
        validate_certs: no
        datacenter: "{{ datacenter_name }}"
      register: gather_all_cluster

    - debug:
        msg:
          - "{{ gather_all_cluster.clusters[cluster_name] }}"

@goneri @Akasurde

There has been an issue with VMware modules not doing URL decoding for some time now, do you want to continue to respond with a workaround?

[RudneiBertolJr]

Hey @sky-joker,

I have tried to test on Ansible 2.9.10, however, the code on the file is a little bit different than the current version on Git.

I imported the unquote lib from ansible and changed the result line to unquote the cluster name, and it worked like a charm.

[sky-joker]

Thank you @RudneiBertolJr for taking the time out of checking the module work :)

[goneri]

Can we get the same problem with some other resource type?

[sky-joker]

Hi @goneri

I think some modules will be the same problem that occurs (mainly the _info-based modules).
I'll find the modules that have the same problem when I have time.

[goneri]

Thank you @sky-joker for the patch.

[sky-joker]

recheck

[sky-joker]

@goneri

Can you please check why the following error occurs by CI?

        "line": 567,
        "column": 4,
        "path": "scripts/inventory/vmware_inventory.py",
        "symbol": "too-complex",
        "message": "'facts_from_proplist' is too complex. The McCabe rating is 22",
        "message-id": "R1260"
    },
    {
        "type": "refactor",
        "module": "inventory.vmware_inventory",
        "obj": "VMWareInventory._process_object_types",
        "line": 690,
        "column": 4,
        "path": "scripts/inventory/vmware_inventory.py",
        "symbol": "too-complex",
        "message": "'_process_object_types' is too complex. The McCabe rating is 28",
        "message-id": "R1260"
    },
    {
        "type": "error",
        "module": "inventory.vmware_inventory",
        "obj": "VMWareInventory.get_instances",
        "line": 352,
        "column": 12,
        "path": "scripts/inventory/vmware_inventory.py",
        "symbol": "assigning-non-slot",
        "message": "Assigning to attribute 'check_hostname' not defined in class slots",
        "message-id": "E0237"
    },
    {
        "type": "error",
        "module": "inventory.vmware_inventory",
        "obj": "VMWareInventory.get_instances",
        "line": 360,
        "column": 12,
        "path": "scripts/inventory/vmware_inventory.py",
        "symbol": "assigning-non-slot",
        "message": "Assigning to attribute 'check_hostname' not defined in class slots",
        "message-id": "E0237"
    }
]�[0m
Cleaning up temporary python directory: /tmp/python-h4xmjwzp-ansible

https://dashboard.zuul.ansible.com/t/ansible/build/b9c74b5e85b44f2c919e876b6b183244

[goneri]

@goneri

Can you please check why the following error occurs by CI?

It should be better now.

[goneri]

recheck

[sky-joker]

recheck

[sky-joker]

recheck

[sky-joker]

recheck

[sky-joker]

recheck