Hi all. I'm noticing some behavior in 10.1.0.4 that seems odd. I have a user A who has been directly granted EXECUTE on DBMS_LOCK and UTL_TCP. Attempts to DROP USER A CASCADE frequently succeed in dropping all of the user's objects, but fail (after the default 5-minute timeout period) to drop the user itself, reporting "timeout waiting to lock object SYS.DBMS_LOCK" (or SYS.UTL_TCP). Now, I understand that grants and revokes must lock the object, and will time out in this manner if the object is in use by another session. I don't recall seeing this, though, in previous versions. I wonder if there's an explicit REVOKE on the DROP USER, or if it's just the recursive SQL to do the revoke. Does any of this ring a bell, i.e., is this the expected behavior even pre-10g? Is there a workaround? Thanks, -- Paul Baumgartel paul.baumgartel@xxxxxxxxxxxx -- //www.freelists.org/webpage/oracle-l