We have CICS regions with the same name running on different LPARs and we would like to monitor these using MainView for CICS. Is this possible?

If not, are there any workarounds we can use?

Updated August 2019: MainView for CICS v6.9.00 introduced the option to use the CICS APPLID to identify the target CICS in Window mode views, instead of the CICS job/stc name.

For v6.9.00 and above, you can use SPATCICS=APPLID in CMRBEXnn to make it easier to have duplicate CICS job names in DIFFERENT LPARs. See the MainView for CICS documentation (Customization Guide) for more details, specifically the section called "Option to use the CICS APPLID as MainView service point identifier"

Please also read the section called "Special requirements and considerations" to understand the implications of choosing this option.

Also note that the MVCICS messages will still refer to the job/stc name, not the Applid. If you need the messages and the CMRDETL history records to match the service point names, using SPATCICS=APPLID is not an option and you will still need to follow the procedures below. 


MainView for CICS v6.8.00 and below

MainView for CICS does NOT support duplicate job names. The job name of the CICS region has to be unique within the PAS Plex (sysplex) for each CICS connected to MVCICS. This has been true since the product was first written. 

If you use duplicate job names - even if they run on different LPARs - unpredictable errors may occur. These include, but are not limited to, the following.

Sporadic connection problems after restarting the CICS
Abends such as S378, caused by attempting to FREEMAIN storage in the 'wrong' CICS
Abends such as S878, caused by PQEs building up in storage as a result of not being able to route them to the correct CICS
Inconsistent CREGION or CREGAGT view output - displaying an active region as INACTIVE, for example
Problems caused by timers, monitors, services, etc., being started in the 'wrong' CICS

There are a few workarounds, as follows, arranged in order of desirability.

1)   You should change your CICS job names to be unique across the LPARs encompassed by the MainView PAS Plex (ie - wherever a MainView CAS runs in the sysplex).

This is the recommended solution.
 

2)   You could change your procedures to use a stepname identifier or qualifier on the START command. The APPLID is the obvious example:

S CICS.cicsappl

Then the service points and communications subtasks - will be created uniquely for each CICS.

Note:  The context names would change to whatever you specified as the identifier on the START command. So, the applid, if you were to use that, would become the context name instead of the job name.

The CT1000D message shows the name of the service point, which is normally the job name:

CT1000D  SERVICE POINT MVCICS  /cicsappl  CONNECT COMPLETED  PAS1

And CREGION would display:

W1 =CREGION==========(ALL======*=======)
CMD Target   SMF    Total       Total  EX
--- Name     ID       CPU  %CPU  Excp  Ra
    cicsappl   EMA1    0.80   0.0  7346  0.

Note: it may also be necessary to change your TCP/IP definitions and any MVS MODIFY commands issued by automation.

3)   You could break the links between the PASes on your different LPARs - BBINODxx nodes and CASDEF definitions - and have them as stand-alone PASes (ie not communicating at all).

Then there would be no problems caused by duplicate job names, but this is a bit extreme. You would have no knowledge of the remote PAS, and the Single System Image power of MainView would be diminished.

1. MainView for CICS