-
Notifications
You must be signed in to change notification settings - Fork 10
Description
Users would like to have control to restart a specific node from the launchpad config, set the prune values appropriately, and run launchpad apply to remove the host. After running launchpad apply, the node that was "reset" still has ucp containers running it. The manager nodes no longer consider the removed host to be in the swarm, but the host still thinks it is a part of a swarm and so keeps all the containers running.
Currently there is a procedure to manually remove the node to leave the swarm, remove all the containers stop running, and then manually clean up the docker installation of images and volumes to prepare for a fresh install, however this procedure is time-consuming and not applicable to the frequent resets required by the users.