Jonathan, thanks for your answers. > One scenario I've created which matches one set of your symptoms: > > Foreign key is deferrable initially deferred. > Session 1 inserts a new parent key value - without commiting. > Session 2 updates a child row changing a value that exists in the parent > table to the new, uncommitted, parent value. > > Session 2 goes into TX mode 4 waiting for session 1 to commit; showing -1 as > the current_obj# I will need to discuss with the developers how exactly the code deals with the inserts, but this will help a lot in that discussion. Maybe it's possible to do all the parent inserts first. Regards Thomas -- //www.freelists.org/webpage/oracle-l