PI77676: CPSM CMAS INCORRECTLY UPDATES MASDIR RECORD RESULTING IN SUBSEQUENT CONNECT FAILURE

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• An issue exists where a CMAS may incorrectly update a MASDIR
  record for the active CICSPLEX. When that occurs, the CMAS is
  subsequently unable to send anything down to the MAS during a
  CONNECT, so a MAS will hang during its connect process, until
  the MASINITTIME time expires. When that occurs, the CMAS EYULOG
  will have a message similar to the following:
  .
  EYUTS0021E CMAS0001 Topology Connect of CICS System MAS00001
                      failed due to a Resource Topology
                      Collection Time-Out.
  EYUTS0002E CMAS0001 Topology Connect for MAS00001 Failed -
                      APPLID(MAS00001) CICSplex(PLEX0001).
  .
  .
  Formatting the trace using  VERBX EYU9D### 'TRC=A,JOB=cmasname'
   you will see exception traces similar to the following:
  .
  Task  Mtd  Prev Tran Obj Level Pt-ID  Debug
  12345 TSSC XLOP TSSC TOP Excp     14 NLGTTOUT
  12345 CTBP CAMM TSSC COM Excp     28 TrgtUnav
  12345 TSSC XLOP TSSC TOP Excp     20 TSSCCPUM
  .
  In the full trace, the MAL for the call to CTBP will show a
  status of SCOPE_NOT_FOUND
  .
  .
  Additional Symptom(s) Search Keyword(s): TSSC EXCEPTION
  METHOD_FAILED TISTTSSC CAMMCTBP
  KIXREVSCB
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICSPlex SM V5R1M0, V5R2M0 and V5R3M0    *
  *                 Users                                        *
  ****************************************************************
  * PROBLEM DESCRIPTION: Communications related problems can     *
  *                      occur when CMASes exchange network      *
  *                      topology information:                   *
  *                                                              *
  *                      -  The communications control block for *
  *                         a MAS (MASDIR) may be updated        *
  *                         invalidly.  This can result in       *
  *                         requests that should be sent to the  *
  *                         MAS to be lost, resulting in time    *
  *                         outs.                                *
  *                                                              *
  *                         If the MASDIR is updated while the   *
  *                         MAS is connecting to its CMAS, then  *
  *                         the connect may fail with variations *
  *                         of the following error messages:     *
  *                                                              *
  *                           EYUTS0021E Topology Connect of     *
  *                                      CICS System <masname>   *
  *                                      failed due to a         *
  *                                      <Resource Topology      *
  *                                      Collection | MAS        *
  *                                      Resource Installation>  *
  *                                      Time-Out.               *
  *                                                              *
  *                           EYUTS0002E Topology Connect for    *
  *                                      <masname> Failed -      *
  *                                      APPLID(<applid>)        *
  *                                      CICSplex(<plexname>).   *
  *                                                              *
  *                      -  MASes connected to a CMAS may not    *
  *                         join to other active CMASes that     *
  *                         manage the CICSplex the MAS is       *
  *                         associated with, or a CMAS may       *
  *                         attempt to join a MAS that is not    *
  *                         active or to which it is already     *
  *                         connected or joined.                 *
  ****************************************************************
  * RECOMMENDATION: After applying the PTF that resolves this    *
  *                 APAR, all CMASes must be restarted.  Note    *
  *                 that the restarts do not need to occur at    *
  *                 the same time.                               *
  ****************************************************************
  When CMASes connect directly or indirectly, they may exchange
  information about their view of the network at the time of the
  connect, including what CICSplexes, CMASes and MASes are active.
  This information usually is routed to all CMASes in the network.
  It is possible that during this information interchange, a CMAS
  will receive some of the information more than once, and for old
  data to be included in the interchange.
  
  Two possible problems that can occur during this interchange
  regarding active MASes have been identified:
  
  -  When information about a MAS is received through this
     interchange, method EYU0CPTI (CPTI) determines whether the
     request is an ADD, UPDATE or DELETE request, and passes that
     information on to method EYU0CPMI (CPMI):
  
     -  if the request is an ADD, CPMI will call method
        EYU0CPAM (CPAM), which will use the provided information
        to build a MASDIR for the MAS.
  
     -  if the request is an UPDATE, CPMI will call method
        EYU0CPUM (CPUM), which will use the provided information
        to update the existing MASDIR for the MAS.
  
     -  if the request is an DELETE, CPMI will call method
        EYU0CPDM (CPDM), which will use the provided information
        to delete the existing MASDIR for the MAS.
  
     The parameter list that CPTI sends to CPMI for an ADD can be
     corrupted, such that it indicates to CPMI that if the CPAM
     call indicates that it did not create a MASDIR because one
     already exists, that CPMI can then call CPUM to update the
     existing MASDIR.  If this occurs and the data that is used to
     update the MASDIR is stale data, including the information
     that this MAS is connected to a different CMAS, then
     subsequent communications requests to the MAS will be
     directed to the other CMAS, where they will be lost.  This
     can result in the problems documented above.
  
  -  If the MAS information received for an ADD indicates the the
     MAS is connected to another CMAS, then CPAM will build an
     EENDJOIN resource table.  This will be delivered to method
     EYU0CPLT (CPLT), which will then check if the EENDJOIN record
     was built from stale data.  It does this by validating that
     the time the MAS connected to its CMAS is later than the time
     its CMAS started.
  
     -  If CPLT determines that the EENDJOIN request is valid,
        then it passes the request onto the Topology component,
        which will initiate a Topology Join request for the MAS.
  
     -  If CPLT determines that the EENDJOIN request is not valid,
        then it calls CPDM to delete the MASDIR that was created
        by CPAM.
  
     A logic error in CPLT can result in the wrong CMAS's start
     time being compared to the MAS connect time.  This can result
     in the problems documented above.
  

Problem conclusion

• To address these problems:
  
  -  CPTI has been updated to ensure it does not corrupt the
     parameter list it passes to CPMI for an ADD MAS request.
  
  -  CPLT has been updated to ensure its uses the start time for
     the CMAS that the MAS is connected to.
  

Temporary fix

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

  PI77829 UI45858 UI45859 UI45860

Modules/Macros

• EYU0CPLT EYU0CPTI
  

Fix information

• Fixed component name

  CICS TS Z/OS V5

• Fixed component ID

  5655Y0400

Applicable component levels

• R00M PSY UI45860

     UP17/03/28 P F703 ¢

• R80M PSY UI45858

     UP17/03/28 P F703 ¢

• R90M PSY UI45859

     UP17/03/28 P F703 ¢

Fix is available

• Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.