PI49270: STATUS NI ON ISRT OF SHISAM INDEX RECORD WITH DEDB SECONDARY INDEX, CI RECLAIM DELETE BYTE VALUE OF X'E0' NOT RECOGNIZED

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• Customer found that when a DEDB root was deleted corresponding
  SHISAM index record was also deleted, but logically, not
  physically. The delete byte was set to x'E0'. This is related to
  CI Reclaim. When the DEDB root was re-inserted, the ISRT of
  the SHISAM index record got 'NI' because the record still
  existed. IMS has logic to detect this situation and use REPL
  instead of ISRT, but the logic was missing in DFSDVSM0 where
  the index processing was done.
  NI is a serious problem because to prevent the indexes getting
  out of sync with DEDB data, FP will not allow a commit after
  an NI on an index update has occurred.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All IMS V13 Fast Path DEDB secondary index   *
  *                 ( FPSI ) users.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: StatusNI occurs on an insert ( ISRT )   *
  *                      of a SHISAM index record using Fast     *
  *                      Path DEDB secondary Index.  The CI      *
  *                      Reclaim delete byte value of x'E0' is   *
  *                      not recognized.                         *
  ****************************************************************
  * RECOMMENDATION: INSTALL CORRECTIVE SERVICE FOR APAR/PTF      *
  ****************************************************************
  When a DEDB root segment is deleted, the corresponding SHISAM
  index record is logically deleted.  Physically the record still
  exists.  The index record delete byte is set to x'E0'.  When
  the root segment is re-inserted, the ISRT of the SHISAM index
  record received a status NI because the record still existed.
  There is logic in DFSDVSM0 to recognize this situation and use
  replace ( REPL ) instead of ISRT, but the logic currently only
  checks for a delete byte value of x'C0'.  A check for delete
  byte value of x'E0' where index processing is done is also
  needed.
  
  Fast Path will not allow a commit after a STATUSNI on an
  index update to prevent the indexes from getting out of sync
  with DEDB data.
  
  Additional Keywords: STATNI NISTAT
  

Problem conclusion

• GEN:
  KEYWORDS:
  
  *** END IMS KEYWORDS ***
  The following module had been modified to correct the reported
  problem:
  
  ************
  * DFSDVSM0 *
  ************
  Code has been added to check the delete byte value of x'E0'
  as well as the existing check of x'C0' delete byte value to
  determine if a REPL of the logically deleted index record should
  be done instead of an ISRT to prevent the statusNI.
  

Temporary fix

• *********
  * HIPER *
  *********
  

Comments

APAR Information

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

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

  PI49287 PI49870 UI32091 PI56194

Modules/Macros

• DFSDVSM0
  

Fix information

• Fixed component name

  IMS V13

• Fixed component ID

  5635A0400

Applicable component levels

• R300 PSY UI32091

     UP15/10/21 P F510 ¢

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.