Closed pLeminoq closed 6 years ago
The problem only occurs when a unit host (a device) is changed which causes its hosted unit to change as well. It was caused because the controller of a device de-/activates its hosted units on configuration changes while the hosted units reinitialize themselves. This alone is not problematic because the whole process is synchronized. But the authenticated communication service overwrote the init method and registered a method outside the synchronized environment. Thus the issue was not a duplicated registration of a method which is filtered, but the registration of a method while the server is already active. The issue should be fixed with the following commits:
Changing the label of a unit during runtime caused an error where the same RPC is registered twice.