Comment 7 for bug 2033980

I _think_ the bug is, that **only** on receiving signal 9 the pid file get's cleaned up:

That is, if radvd crashes for example, the pid file does not get cleaned up, at least when I search neutrons codebase for the function I can't find any other code that calls it in the context of radvd.

So, if radvd does not clean up after itself properly when crashing - which just might not be possible - I think this is a bug in neutron, but it's possible I have some misunderstanding, because this involves multiple layers of process management (python, Neutron ProcessManager, eventlet..).