What are the recommended steps to take when the docker daemon needs to be stopped on a managed host?
It probably depends on what you are expecting to happen with your services. If you have any services that require no down time, then you’d want to make sure you have another host that has the services running.
Also, depends on if your services have health checks, if you have these health checks on, you could deactivate a host and have your services move to an active host before restarting.
well, my intent was to find out if there’s something we should/could do to ensure the host reconnects to the server when it comes back online. This was triggered by what seems now to have been a one-off or some other error. We had to restart one node recently, and it wouldn’t re-register on its own. We ended up having to delete the host from the server and re-add it manually, which was quite a pain.
More recently, though, we had a complete outage of all of our nodes and they came back up without any steps being taken on our part, which is more what I expected.
I suppose if that’s the expected behavior, then anything else is just troubleshooting if we have problems in the future.