Вот, сегодня в эхе прочитал. Посвящается всем, кто ещё не замучил ошибку 10038 :).

The error code 10038 is WSAENOTSOCK (see http://tinyurl.com/gth2 [shortcut to microsoft.com])

This means that for some reason the socket handle the engine has been using all of a sudden has become "bad". I cannot think of anything that would cause a socket to go bad, short of a network or hardware glitch (e.g. flaky NIC).

You might try running your database engine on a different NIC or a new machine, just to rule out the possibility of local hardware issues.

There is a new (hidden) option that was added to ASA 8.0.3 (5181) [and newer 9.0.x builds too] specifically to let the engine continue while ignoring socket errors (there are some cases where Winsock throws incorrect or ambiguous errors/warnings). Simply add the following to your server startup options:

-x tcpip(IgnoreCodes=10038)

and the engine will ignore any 10038 errors it gets.
HOWEVER...recognize that ignoring the error may put the engine into a bad state(!!)

I suggest trying this on a test database where you have been able to reproduce the 10038 error condition. See if adding the switch lets the engine continue to accept new connections past the point where it had been closing the listener.

If you still are facing problems, you might consider opening a case with iAnywhere Technical Support.

Hope this helps,
greg.fenton
--
Greg Fenton
Consultant, Solution Services, iAnywhere Solutions
____________________________________
- Гарфилд, мышь!
- Спасибо, я сыт!