Please see the description of my issue and my follow-up comments here https://gadgeteer.codeplex.com/workitem/2698.
I thought I had a work around, but apparently it’s still lacking.
I do not have an answer but now your hidden link is revealed.
Aha, I read his name and assumed somehow he made the link hidden but it was just because he was a newbie
Any word? Has Windows 10 on Raspberry Pi 2 distracted everyone? 
http://www.hanselman.com/blog/SettingUpWindows10ForIoTOnYourRaspberryPi2.aspx
I’m pretty sure people are doing their own things - some of them certainly are RPi2, but not everyone
Anyway, I didn’t understand your last comment on codeplex. The cable gets removed, then you make a request to the server before the stop() happens ? In that scenario the request should never hit the device because the cable is out ?!?
- Ethernet Cable Connected
- Web Server Started
- HTTP Client Request
- HTTP Server Response
- Ethernet Cable Disconnected
- Web Server Unstoppable
- Wait ~10 seconds
- Ethernet Cable Connected
- Web Server Start Fails
The part that threw me off was that steps 3 and 4 have to happen - a client must make an HTTP request to the web server. If steps 3 and 4 do not happen, I’m able to stop the web server (step 6) properly, so everything works. It seems as if adding steps 3 and 4 causes a thread to be spun up such that the socket cannot be closed?
Since nobody responded, I posted my unacceptable, hacky work around in the comments section of the codeplex page which I referenced earlier in this thread.