Garrick, I think I found it. It was a buffer overflow in the node allocation routine which was only triggered with jobs over 500 tasks with PBSLOGLEVEL set to 3 or higher. It is fixed and a new release rolled. Dave