If the system detects an unrecoverable error and abnormally ends
the APPC address space, and APPC automatically restarts, message ATB005I
indicates that APPC is restarting. In addition, the following occurs:
- The cross-system coupling facility (XCF) notifies the members
of APPC's XCF group that APPC is unavailable.
- The system shuts down all active logical units (LU).
- The system does not save or restore active APPC conversations,
either local and remote. The conversations will be lost and the system
will notify the invoking applications.
- The system will not read the APPCPMxx parmlib member during restart.
The parmlib members that were in effect prior to the restart will
still be in effect following the restart.
- The system deletes the cache of TP profiles and side information.
Following restart, the system refills the cache during normal processing.
- If APPC component trace is active before APPC abnormally ends,
the system dumps the trace records when APPC abnormally ends and the
APPC trace will not be active following restart.
- The system does not save or restore information regarding schedulers
when APPC abnormally ends and restarts.
- XCF notifies the members of the APPC's XCF group that APPC is
available once it restarts.
Recovery Actions for APPC Abnormal Ending with Restart
- Have applications redrive the system data file manager (SDFM)
activate/deactivate TP profile interface to reset the TP profile information
following the restart.
- Restart APPC component trace and all APPC conversations that were
active before the error when APPC services are again available.
- Have schedulers re-identify themselves after APPC has restarted.