During
normal task termination, a resource manager might
end abnormally; its own recovery routines, if any exist, will receive
control. If they do not retry, or if the resource manager has no
recovery routines, the system now considers this situation to be an
abnormal termination,
and passes control to the newest ESTAI routine. Because the abending
resource manager, and any previous resource managers, might have completed
some processing, the ESTAI routine will run in an unpredictable environment.
In this situation, IBM® recommends that you restrict
the ESTAI routine's processing. For the ESTAI routine to run in this
environment, design it to:
- Check the STCBRMET field in the STCB; if the bit is on, the ESTAI
routine is running after a resource manager has ended abnormally
and its recovery routines have not retried. In this situation, the
ESTAI routine does not need to hold a lock to check the STCBRMET field.
For the mapping of the STCB, see z/OS MVS Data Areas in
the z/OS Internet library.
- Do as little processing as possible, and nothing that might depend
on a resource that might have been cleaned up already.
- Do not request to retry. The system will not allow a retry
in this situation.
Note that no other ESTAE-type routines receive control in this
situation; only those established through the ATTACHX macro still
exist at this point in termination processing.