Table 1 helps you make sure that the
problem is in RACF®, rather
than in either the caller of RACF or
a system service used during processing of a RACF request.
For some problems, you might be able to do a quick check of the
problem symptoms in the dump (if present), the job and system message
log, SYS1.LOGREC, trace records, or other problem output.
Table 1. Determining Whether
You Have a RACF ProblemQuestions |
Recommendations |
---|
Was there a valid request
for a RACF function? |
This could be that: - A RACF macro was invoked
- A RACF utility was invoked
- A RACF command was invoked
- A user selected a RACF option
on an ISPF panel
The RACF request must
be valid. A user request could invoke a routine that in turn invokes RACF. A RACF error message (like ICH408I, insufficient
access authority) would then be issued not because of a RACF error or because of an error made by the
original user, but because of the routine that invoked RACF.
Note: If there is a system dump, you
can check this in the trace records recorded in the dump. If there
is no system dump, you may need to recreate the problem and request
a system dump when the problem recurs.
|
Were error messages issued by system
services that were used during RACF processing? |
Check whether any error messages
were issued for the user request, and make sure that message IDs are
included with the messages. If a batch job produced the output,
check that the job statement had MSGLEVEL=(1,1) specified. If a TSO/E
user had the problem, check that the user profile had these options
specified: PROFILE WTPMSG MSGID.
|
Did RACF return
control to the program that made the request? |
For RACF macros
and utilities: if there is a system dump, you can check this in the
trace records recorded in the dump or (if an SVC was issued) in the
RB chain in the system dump. If there is no system dump, you may need
to recreate the problem and request a system dump when the problem
recurs. For batch jobs, you can also check whether the job step for
the program completed with a zero return code. (If an application
program made the request, the TSO TEST command, a SLIP command issued
by the system operator, or a WTO or PTRACE macro added to the program
could be helpful in checking the execution steps within the program.) |
Note: While diagnosing the problem, you need to continue to check
whether something other than RACF might
be the source of the problem. Possible candidates are:
- Callers of RACF
- System services used by RACF
- A hardware error
- A program that overlaid storage used by RACF, its caller, or a system service
- A storage shortage can cause RACF commands
to ABEND unpredictably