tralafiti

I noticed that after stopping a child it's PID was still listed in contexts children map. Extended test case and fixed the bug

When poisoning a non-existing (e.g. already crashed) pid, the waitGroup would deadlock. Added test and fixed by only incrementing waitGroup if the process was found.

Host network container unreachable after restart

1

### Description When OS or Docker Desktop is restarted, container with host network are not reachable anymore. ### Reproduce Create container with host-network and restart policy ```bash % docker run...