extproc stops listener from unloading properly

We have started replacing some external job servers with modules that are called via the external proc facility in 805.

During testing we have noticed that with some DLL once they are called we have problems stopping and restarting the listener.

Specifically after we stop the listener when we try to restart it fails basically saying that address is alread in use and hinting that the problem is that a socket didnt disconnect when we stopped the listener. The only way we can then restart the listener is to shutdown the whole db.......

On a test server this is annoying but bearable but it worries me when we bring this up on a live server and get the same problem.

We are running 805 on NT4.

I can understand a proc trash the listener (which is why we might stop the listener but there are other reasons of course) but I am surprised that the db itself can get locked up...

Is this known behaviour, is there a fix or a workaround????

Rob Received on Tue Jan 26 1999 - 07:43:58 CST