Warning: don't mess with rolling restart

I apparently put myself in a pickle last night trying to get "fancy" with the status of services after a rolling restart. Since I was doing several maintenance tasks while also trying to minimize downtime for Zope users I had shut down things that didn't need to attempt to run before I would be shutting them down again (e.g. Analytics export, CMDB sync, collection in general, etc.).

Once I finished a ZenPack install I did a rolling restart...then tried to immediately shut down those things above in the UI. Then I moved to the next thing...and my request to take a snapshot hung. I went back to the UI and the things I had tried to shut down were now started, so I shut them down again...well, then it got worse. Everything else (even stuff that wasn't part of the rolling restart) went to "paused" status, and then I learned that when things are paused you don't tell everything to start (managed to lock both MariaDB/MySQL DBs).

But the big lesson: Let the rolling restart finish before you mess with stopping services. Oh, and by the way, clearing the locks on MySQL I just OS restarted my resource pool nodes...I tried to review with docker ps -a, but couldn't find any containers hanging out, but that could have been a timing thing. Historical HBase container metrics did indicate that CC did start a 2nd container before the first one was shut down for several containers that I saw (basically showing counts that were too high to be just one or one set of containers).