Diagnosing a CQS related problem
CQS produces SDUMPs for internal errors. The CQS dumps are in the SYS1.DUMP data sets. CQS can also produce LOGREC data set entries for errors.
CQS related problems
For a CQS environment, related problems might include:
- IMS WAIT problems
- CQS WAIT/HANG problems
- CQS checkpoint problems
- CQS restart problems
- CQS structure rebuild problems
Implement normal operating procedures to preserve the following documentation close to the time the error occurred:
- Additional manual dump intervention
- z/OS® log stream (for problems related to IMS shared queues)
- Most recent SRDS (structure recovery data set) for each dumped structure
For a CQS WAIT/HANG problem, obtain dumps and syslogs of CQS address spaces in the sysplex such as:
- One dump and syslog from the master CQS
- One dump and syslog from the non-master CQS
- One dump and syslog from the error CQS
- One dump and syslog from the normal CQS
A sysplex contains only one master CQS and, most likely, one error CQS in a sysplex. Thus, the maximum number of CQS dumps and syslogs taken is three. If the sysplex contains fewer than three CQS address spaces, then dumps and syslogs of all CQS address spaces are needed.
Before obtaining the syslog, issue the following z/OS DISPLAY
commands
to write the sysplex information to the syslog:
D CF,CFNAME=cfname
D XCF,CF
D XCF,STRUCTURE,STRNAME=strname
For a CQS loop problem, obtain two dumps. To obtain the two dumps:
- Obtain a z/OS SVC dump
of the CQS and its associated IMS control
region address space by issuing the following command:
DUMP COMM=(dump title) R id,JOBNAME=(j1,j2), SDATA=(CSA,PSA,RGN,SQA,SUM,TRT),END
In this command, j1 is the CQS job name, and j2 is the IMS control region job name.
- Save the IMS log data sets that are created during the error period.
- Save the current z/OS log data sets that are created. You can copy the current z/OS log data sets for the CQS log stream by using the IEBGENER utility. No z/OS log data sets are logged (unlike the IMS logger, which does have log archive capability through SLDS).
If an isolated event type within CQS encounters an error, IBM® Software Support might request additional trace level settings for the various trace types. If a structure rebuild or structure checkpoint related problem occurs, you will also need to dump the CQS address spaces for any CQS associated with the given structure, and save the associated SRDS (structure recovery data set) for the CQS structure checkpoints and CQS system checkpoints.