I have a C# windows service using UDP client to send/receive UDP records for a billing accounting application. At some point, a connectivity problem arises and even a manual ping fails.... later the ping succeeds and my service still can't send/receive packets with the far end. The only thing that resolves the issue is restarting the service. Closing and destroying the UdpClient object doesn't seem to work. Also, the protocol I am using has "heartbeat" packets that we send on an interval when things begin to time out ... those also cease to make it to the far end when this condition occurs.
Is there something cached in UDP client and/or the .NET framework that causes the entire process to be terminated to recover? I do believe after many attempts to diagnose using wireshark, logging, etc. that there must be some stale routing information cached in my process causing this problem. I'd rather not resort to spawning a second process to restart the windows service. That would be ugly and unnecessary in my opinion.
Any help would be appreciated.