Ensure no zombie processes are left over by run.sh

[rodja]

This PR applies some additional checks to ensure we do not create zombie processes. In one of our projects we saw many of these:

message+  725270  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [exim4] <defunct>
message+  725271  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [exim4] <defunct>
root      725282  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [sudo] <defunct>
root      725285  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [sudo] <defunct>
message+  725288  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [exim4] <defunct>
message+  725289  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [exim4] <defunct>
root      725297  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [sudo] <defunct>
root      725300  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [sudo] <defunct>
message+  725303  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [exim4] <defunct>
message+  725304  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [exim4] <defunct>
root      725309  0.0  0.0      0     0 ?        Z    Okt19   0:00 [sudo] <defunct>
root      725311  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [sudo] <defunct>
root      725314  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [sudo] <defunct>
root      725316  0.0  0.0      0     0 ?        Z    Okt19   0:00 [arp-scan] <defunct>
message+  725317  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [exim4] <defunct>
message+  725318  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [exim4] <defunct>
root      725326  0.0  0.0      0     0 ?        Z    Okt19   0:00 [sudo] <defunct>
root      725328  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [sudo] <defunct>
root      725331  0.0  0.0      0     0 ?        Zs   Okt19   0:00 [sudo] <defunct>
root      725333  0.0  0.0      0     0 ?        Z    Okt19   0:00 [arp-scan] <defunct>

Which leads to BlockingIOError: [Errno 11] Resource temporarily unavailable down the road (when Linux can't create any more processes).

[rodja]

It's still happening. So I need to investigate it further tomorrow.

[rodja]

After many general improvements to run.sh I still got a lot of zombie processes. Then I discovered that calling a command with sudo in a docker container creates a zombie process. This is fixed with cc72d89 in this PR. But still I got a slow increasing count of zombie processes. After some digging I saw that it was caused by SIGKILL as abort signal for the timeout (fixed in 3024e48).