Recommended documentation
- Dumps of the failure
- System generalized trace facility (GTF) trace output
- SMPLOG data
Follow this procedure to gather the recommended documentation and
report the problem.
Attention: SMP/E can appear to loop for a long time, depending
on what it is doing and what system it is running on. This does not
necessarily indicate a problem. Do not cancel an SMP/E job unless
you are sure it is looping or in a wait state. Canceling a job could
damage the SMPCSI or the other data sets SMP/E uses. If the job seems
to be looping or waiting too long, call your IBM® Support Center to see if you should cancel
it.
- If you have an online I/O checker, such as resource measurement
facility (RMF™), use it to see
if:
- No I/O is being performed.
- The same number of execute channel programs (EXCPs) is used continuously.
- Check the SMPLOG to see if:
- No messages are being printed.
- The same message or group of messages is printed over and over.
- Slip-trace the branch and link register commands (BALRs) in the
system GTF trace, and look for repeated BALRs or repeated strings
of BALRs between SMP/E modules.
- If you have a backup copy of SMP/E, its data sets, and your system
data sets, try to recreate the problem. Then, take several dumps of
the failure to trace a consistent PSW instruction address or range
of addresses. To take the dumps, use a SYSABEND DD card and cancel
the job that contains the problem.
- Report the problem to the IBM Support
Center.
After asking for your account name, license number, and
other customer identification, the service representative will ask
for a brief description of the problem.