Some IRC chatter highlighted an issue where the logs of a dying controller - and I suspect a dying model as well, being the root cause - are spammed with:
2019-05-22 21:35:19 ERROR juju.worker.modelcache worker.go:171 watcher error, shared state watcher was stopped, getting new watcher
2019-05-22 21:35:19 ERROR juju.worker.modelcache worker.go:171 watcher error, shared state watcher was stopped, getting new watcher
2019-05-22 21:35:19 ERROR juju.worker.modelcache worker.go:171 watcher error, shared state watcher was stopped, getting new watcher
2019-05-22 21:35:19 ERROR juju.worker.modelcache worker.go:171 watcher error, shared state watcher was stopped, getting new watcher
It seem the model cache worker is not respecting the dying state and attempts to continually, and fruitlessly - restart the watcher. It also seems to result in a stuck dying operation, but not 100% sure it does cause that.
I have addressed this on 2.6 (https:/
What version of Juju did you see this on?