Idempotency Cleanup: [NODES]

[tbielawa]

You really should run ara to make your life easier fixing up these items. Sharing the results is difficult due to the size of the generated static reports. Instead below you'll find a copy/paste of tasks and a screenshot.

There are batches of tasks that show up as changed every time the upgrade playbook is ran. These tasks are only from the nodes list of changed tasks. These require review:

Clean package cache	n01.example.com	command	0:00:14	0:00:03	CHANGED	details
Mark node unschedulable	n01.example.com	oadm_manage_node	0:06:21	0:00:00	CHANGED	details
Drain Node for Kubelet upgrade	n01.example.com	command	0:06:22	0:00:01	CHANGED	details
Restart docker	n01.example.com	service	0:06:59	0:00:01	CHANGED	details
Restart rpm node service	n01.example.com	service	0:07:02	0:00:01	CHANGED	details
Set node schedulability	n01.example.com	oadm_manage_node	0:07:04	0:00:00	CHANGED	details

[sdodson]

I think these are all playbook level idempotency issues. I'm not sure how we can address them, we have to decide if and when it makes sense to completely skip a node. Perhaps if we have verified that its version already matches the desired version?

I think these can all get a pass for now.