FEPI does not (deliberately) issue either CICS® transaction abends or MVS™ abends. However, an unexpected failure can occur in the following places:
These abends have different results, as shown in Table 10.
Point of failure | Result |
---|---|
Application | The usual transaction abend for the error condition. |
EXEC stubs | The usual transaction abend for a failure within CICS management modules. An example of this is an ‘operation’ program check, which generates a CICS AKEA abend, which in turn generates an ASRA abend. |
FEPI adapter | The usual transaction abend for a failure within CICS management modules. An example of this is an ‘operation’ program check, which generates a CICS AKEA abend, which in turn generates an ASRA abend. |
FEPI Resource Manager | No direct effect on the application program, because the abend occurs under the CSZI Resource Manager task. This probably results in a DFHSZ4099E message (see Message DFHSZ4099E), and the failure of the Resource Manager. An example of this is an ‘operation’ program check, which generates a CICS AKEA abend, which in turn generates an ASRA abend. Any CICS FEPI transactions are left waiting on the FEPI_RQE resource (for details of FEPI waits, see the CICS Problem Determination Guide). |
VTAM exit | A VTAM abend; a VTAM dump is taken. Because the exit lies within the FEPI Resource Manager, the CICS abend handling routines are activated to process a "normal" failure in the Resource Manager. |
An abend in an application program, an EXEC stub, or the FEPI adapter affects only the active CICS task that issued the FEPI command; other FEPI programs continue as normal.
If an abend affects the SZ TCB, CICS makes that TCB unavailable for use, while keeping the other CICS TCBs active and accessible. This means that FEPI functions can be restored only by restarting the CICS system.
This message indicates that the abend exit routine within the FEPI adapter has trapped an abend within the FEPI Resource Manager.
As soon as an abend within the Resource Manager is detected, the FEPI state (in the FEPI static area) is set to ‘Failed’. If possible, message DFHSZ4099E is issued, together with a SNAP dump, to indicate that FEPI has failed. However, in some circumstances it is not possible to issue DFHSZ4099E, and a system dump is generated instead.
Any FEPI transactions are left waiting on the FEPI_RQE resource (for details of FEPI waits, see the CICS Problem Determination Guide). These waits never get posted, so the transactions suspend. You must issue a CEMT FORCEPURGE command to remove these suspended transactions from the system.
This message indicates that a connection has ended, and gives a reason code taken from the VTAM control blocks. The reason code may be returned in the LASTACQCODE option of a CEMT or FEPI INQUIRE command, depending on the operation which generated DFHSZ4155I.
DFHSZ4155I does not always indicate a problem; if you took positive action to end the connection, DFHSZ4155I merely confirms that VTAM did as you requested. However, if the connection ended unexpectedly, the reason code tells you why.
To determine what the reason code means, refer to the VTAM Programming manual.
[[ Contents Previous Page | Next Page Index ]]