drastic sleeping forevers lock the database.

Hi, one of my SPL-programs encountered a few times a "sleeping forever" condition that looked more like some kind of deadlock (the process would never awake). Now, the problem is that as soon as I "trace" all instructions, no sleeping forever-deadlocks are encountered any more... very strange... Killing the process does not help (the process remains asleep and prevents others to access this database, we need to reboot the server when this happens). Pure from a programming point of view : what SPL-instruction combinations can cause the system to go into such a drastic "sleeping forever" ???? The program works extensively with small temp tables (no indexes because of small size) that are created and dropped every time again. Performance is avg 5 times per second, but I think performance can be increased dramatically when we find out what causes the "sleeping forever" conditions. System : Compaq NT-server, Ifmx v9 Kind regards, David Van Huffel.