Unnecessary Keystone dependency in resource agents

[Norbert Illes]

The monitoring function of some resource agents are depending on the availability of Keystone: they try to get a token from Keystone and they are using this token during monitoring.
If Keystone gets unavailable, these Keystone-dependent resource agents mark these -otherwise running- resources unavailable, simply because they can't get a token from Keystone, causing unnecessary resource restart/migration.

Although, we shouldn't consider the resources fully available if Keystone is not there (e.g. if neutron-server is running but Keystone isn't, the user can not use Neutron properly), this dependency should not be hard-coded in the resource agents.

The impacted resource agents:
 * neutron-server
 * glance-registry
 * manila-api
 * cinder-api
 * nova-api

[Adam Spiers (adam.spiers)]

Norbert filed https://review.openstack.org/#/c/275108/ to fix this but unfortunately IIUC it seems it referenced the wrong bug number. @Norbert, can we mark this as resolved now, or is there a reason your commit message said only "Partial-Bug" ?

[Norbert Illes (nilles)]

The reason I put "Partial-Bug" on https://review.openstack.org/#/c/275108/ is that, it only removes Keystone dependency from neutron-server, and there are more resource agents impacted by this bug, as it stated in the bug description.
If it's ok with you, I would only consider this bug closed after the Keystone dependency removed from all of the mentioned resource agents.

Regarding the wrong bug reference:
It probably happened because I accidentally included a wrong bug number in my first patch set. I corrected this in the second patch set, but it seems the linking between Gerrit and Launchpad already happened by that time, so "OpenStack Infra (hudson-openstack)" put all of its comments to the wrong bug:
Comments #3 and #4 here: https://bugs.launchpad.net/openstack-resource-agents/+bug/1511721 should be here.

[Adam Spiers (adam.spiers)]

Yep, that's fine by me. Thanks for the explanation!

And don't worry about the wrong bug reference. I make mistakes too ;-)