Because Exit 12 is called from every stage in JES2 processing, there are significant variations the recovery environments JES2 provides for Exit 12. For example, when $STRAK is called from HASPRDR, an error in your exit routine may cause only the current job to fail; however, when $STRAK is called from HASPNET, an error in your exit routine may cause JES2 itself to fail. As with every exit, your exit routine should not depend on JES2 for recovery. JES2 cannot anticipate the exact purpose of your exit routine, and therefore any standard JES2 recovery that happens to be in effect is, typically, minimal. You should provide your own recovery within your exit routine.