Resetting a running environment sends a close_command to the simulation. Then, the ns3 simulator is destroyed, and all ScheduleDestroy events should be handled:

In parallel, the experiment is runned again, which kills the old environment and interrupts the ScheduleDestroy events.

Instead, I expected that all events scheduled with ScheduleDestroy are executed when I reset the environment. It would be a nice feature to execute these events, when I close the environment.