Tower CUD check and run refresh_in_provider followed by refreshing manager

[jameswnl]

The targeted refresh is only implemented on confirguration_script_source and so if it is in the context of updating a credential, the refresh on the credential will be dropped.

In prior PR, was convinced by this section that it will fall back to ems refresh automatically.

Now realized, it will be dropped by this queue attempt way before.

@miq-bot add_labels fine/yess, providers/ansible_tower, blocker, bug

[miq-bot]

@jameswnl Cannot apply the following label because they are not recognized: fine/yess

[jameswnl]

@miq-bot add_labels fine/yes

[miq-bot]

Checked commit jameswnl@ee90bbf with ruby 2.2.6, rubocop 0.47.1, and haml-lint 0.20.0
2 files checked, 0 offenses detected
Everything looks fine. 👍

[blomquisg]

The targeted refresh is only implemented on confirguration_script_source and so if it is in the context of updating a credential, the refresh on the credential will be dropped.

@Ladas, can you take a look at the graph refresh flow? Need to make sure that if a targeted refresh is requested, but not implemented, then the full refresh is kicked off instead. Can be done as a separate follow-up PR. But, I don't think the callers should care if the targeted refresh exists for specific objects yet. The callers should be able to assume it can call a refresh with a target and the target gets refreshed (even if it means it's a full refresh).

[Ladas]

@blomquisg right, we need to redesign the target refresh queuing, this queues the AR object, so it will be filtered out by the core code. @durandom ^

[simaishi]

Fine backport details:

$ git log -1
commit 647e49c22715dd11a53011de281f27e0be5f0ba0
Author: Greg Blomquist <[email protected]>
Date:   Mon May 8 17:26:49 2017 -0400

    Merge pull request #15025 from jameswnl/avoid-target-refresh
    
    Tower CUD check and run refresh_in_provider followed by refreshing manager
    (cherry picked from commit 2b14633149392a51c11490555f5f2ecb27bd5794)
    
    https://bugzilla.redhat.com/show_bug.cgi?id=1448098