Could the locking issue here by that the transacion self-locks? Scenario: the update has a write lock, and then the select count(*) called from the trigger's stored procedure can't get a read lock because Derby deoesn't recognize this is the same transaction?
Dag
