Identifying a problem

Before you can determine the cause of a problem, you need to collect as much information as you can about your system and the symptoms you are experiencing. This topic describes some preliminary checks for problem determination.

The following basic questions help you identify important information. As you go through these questions, make a note of any changes to your environment and of any unusual occurrences, regardless of whether you think they are relevant. Even if the conditions you observe do not at first appear related to the problem, information about them could be useful later if you have to carry out systematic problem determination.

Has CICSPlex® SM run successfully before?

If CICSPlex SM has not run successfully before, it is possible that the system has not been installed or set up correctly.

For information about installation and setup requirements, refer to the Program Directory (or other installation instructions). In particular, you might want to try running the installation verification procedures (IVPs), which are described in Verifying the CICS installation. These procedures are designed to verify the correct installation of CICSPlex SM libraries and components.

Have any changes been made since the last successful run?

If CICSPlex SM has run successfully in the past, review any changes that have been made to your data processing environment since that time.

Think® about your operating systems, CICSPlex SM itself, the CICS® systems it manages, the hardware they run on, and any related operational procedures.
  • If an APAR or PTF was applied to any of your operating systems, CICS, or CICSPlex SM, check for error messages related to the installation. Also check for any unresolved ++HOLD ACTION items associated with the SMP/E maintenance. If the installation of maintenance was successful, check with IBM® Support for any known APAR or PTF error.
  • If a hardware modification was made, it may have affected the systems on which CICSPlex SM runs or the connectivity between systems in a CICSplex.
  • If your initialization procedures changed, check for messages sent to the console during CICSPlex SM or CICS initialization. It could be that changes to JCL, CICS system initialization parameters, or CICSPlex SM system parameters are causing a problem.
  • If the configuration of one or more CICSplexes has changed, check the EYULOG consolidated message log for messages describing incorrect or incompatible definitions. For example, if you are migrating additional CICS systems to management by CICSPlex SM, ensure that the topology definitions for the new systems have been added to the CICSplex.
Are there any messages that could explain the problem?

Check to see if there were any unusual messages issued during CICSPlex SM initialization or immediately before the problem occurred.

Also check for any messages related to a CICS system that is being managed by CICSPlex SM.

If you find any messages that you don't understand, refer to the appropriate messages manual for an explanation and a recommended course of action.

Does the problem occur at specific times?

If the problem seems to occur only at specific times of the day, consider what's happening in the system at that time.

  • How many MASs are active? Where are they located and how are they communicating with the CMAS that is managing them? Have any MASs or CMASs recently become active and begun communicating with other address spaces?
  • Could the problem be related to system loading? Is the number of MASs (with associated resource activity) at its peak? If your CICSPlex SM environment extends across more than one time zone, remember that the time of peak system usage may vary.
  • What type of monitoring, workload management, or analysis definitions are in effect? Keep in mind that the use of time periods can cause definitions to automatically become active or inactive at preselected times of the day.
Does the problem affect specific parts of the environment?

If the problem seems to affect only certain parts of the CICSPlex SM environment, consider what is unique about those parts.

If, for example, just one CMAS is experiencing a problem, review its configuration definitions:
  • What system parameters were used in its startup job?
  • What other CMASs does it communicate with?
  • What CICSplexes does it participate in managing?

Common types of problem

Common problems include abends, stalls, bottlenecks, or problems with CICSPlex SM system management functions.

Here is a summary of the common problems:
  • An abend has occurred.

    CICSPlex SM-generated console, job log, or TSO terminal messages indicate that an abend occurred and provide an abend summary.

  • A stall has occurred.

    The system is not responding to users logged on (to the MAS), or the system is using an abnormally low number of processor cycles or no processor cycles.

  • A bottleneck has occurred.

    The system response is abnormally slow, or the system is using an abnormally high number of processor cycles.

  • CICSPlex SM system management functions are not working as expected.

    For example, monitor or analysis definitions are not active, real-time analysis events are not occurring or are not being resolved, or a workload is being routed incorrectly.

For more details, see Investigating output and system management problems.