Troubleshooting data for CICSPlex SM problems

CICS® MustGather for CICSPlex SM problems

Gather the following diagnostic information before contacting the CICS support team to troubleshoot your CICSPlex SM problems.

Required data:

The complete CICS job output for the CMAS and the MAS or WUI server that experience the problem.
For problems that are related to a CMAS, in addition to the output for the CMAS experiencing the problem, gather the output from all other CMASs in the CICSplex. The requirement for gathering documentation from all CMASs is due to the way CMAS to CMAS communications works.

For problems that are related to a MAS or WUI server region, gather the output for the MAS or WUI server and for the CMAS to which the MAS or WUI server connects.

This output must include the job log, JES messages, JCL, MSGUSR DD output, and the EYULOG DD output. Do not limit the size of the output by excluding anything. CICSPlex SM problems are often caused by events that occurred minutes, hours, or even days before you notice anything unusual.
The complete MVS™ system log from the LPAR where the problem occurs. Do not shorten or cut down the size of the log in any way. If you attempt to cut out and send only what you think is relevant, you might inadvertently cut out critical information.
The MVS system dump that is taken at the point of failure. For more information, see Using dumps for CICS problem determination on z/OS®.
If your CICS region hangs or loops and you do not get a CICS or CICSPlex SM dump or message when problem occurs, dump the CICSPlex SM environment with the IEADMCCP command. This command is supplied with SYS1.SAMPLIB. For details, see IEADMCxx (DUMP command parmlib) in z/OS MVS Initialization and Tuning Reference. For more information about dumping CICSPlex SM, see Using the MVS DUMP command. If you do not automatically get a dump but receive a CICS or CICSPlex SM message when the problem occurs, set an MVS SLIP trap to take an SVC dump of the appropriate address spaces and data spaces.

For example, if you always get an EYUXL0105E message in your MAS when a certain problem occurs, set your SLIP trap as follows:
```
SL SET,MSGID=EYUXL0105E,ID=CPSM,ACTION=SVCD,MATCHLIM=1,
R y,JL=(cmas,mas,EYUX###),DSPNAME=('EYUX###'.*),
R y,SDATA=(ALLNUC,LPA,LSQA,PSA,RGN,SQA,TRT,CSA,GRSQ),END
```
y

This parameter specifies the reply identifier that displays on the console.

cmas

This parameter specifies the name of the CMAS that the MAS connects to.

mas

This parameter specifies the job name for the region you see the problem message.

EYUX###
In both the JL and DSPNAME parameters, this parameter specifies the ESSS address space and must reflect the release of CICS that you are running:
- For CICS TS 4.2, use EYUX420 in both parameters.
- For CICS TS 5.1, use EYUX510 in both parameters.
- For CICS TS 5.2, use EYUX520 in both parameters.
- For CICS TS 5.3, use EYUX530 in both parameters.
- For CICS TS 5.4, use EYUX540 in both parameters.
- For CICS TS 5.5, use EYUX550 in both parameters.
Ensure that you dump the CMAS and EYUX### address spaces even if you see a problem only in one MAS.
If you take a dump by using the MVS console DUMP command, make sure that you take a dump of the ESSS and its data spaces too. The following example is a sample DUMP command that you can use:
```
DUMP COMM=(DumpTitle) 
R xxx,JOBNAME=(cmas,mas1,mas2,EYUX###),DSPNAME=('EYUX###'.*),CONT
R xxx,SDATA=(ALLNUC,LPA,LSQA,PSA,RGN,SQA,TRT,CSA,GRSQ),END
```
You can add extra MAS names to the JOBNAME parameter as shown in the example above.
Unformatted (raw) CICS auxiliary trace data sets (DFHAUXT and DFHBUXT) from the CMAS and the MAS or WUI server. For more information, see Using trace for CICS problem determination on z/OS.
For problems that are related to a CMAS, in addition to the unformatted (raw) AUX trace files for the CMAS, gather the unformatted (raw) AUX trace files from all other CMASs in the CICSplex. The requirement for gathering documentation from all CMASs is because of the way CMAS to CMAS communications works.
When you have a routing failure and you want to know why, capture the CICSPlex SM WLM trace that is generated from the routing region. Also, include the dump and the auxiliary trace from both the CMAS and the routing region. For more information, see Using trace for CICS problem determination on z/OS.