PK48044: ABEND S0C4 IN MODULE FRXQBUF0 AT OFFSET = 00000698 BECAUSE AWRVNRD_RVQB IS ZERO

APAR status

• Closed as program error.

Error description

• ABEND S0C4 in module FRXQBUF0 at OFFSET = 00000698 because
  AWRVNRD_RVQB is zero
  AWRVFNRD--All bytes contain X'00'
  An AWE was queued from SEQ#25231800 in FRXRDTH0, with
  awrvnrd_rvqb <= awrvnrd_queue_token = rvra_queue_token = 0
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All DRF R310 users might be affected.        *
  ****************************************************************
  * PROBLEM DESCRIPTION: ABENDS0C4 in FRXQBUF0 after log         *
  *                      read failure.                           *
  ****************************************************************
  * RECOMMENDATION: INSTALL CORRECTIVE SERVICE FOR APAR/PTF      *
  ****************************************************************
  If a read error for a log data set is detected by DRF and a
  copy of the log exists (such as a secondary RLDS or primary
  SLDS), an ABENDS0C4 in FRXQBUF0 can occur.
  

Problem conclusion

• AIDS: RIDS/UTIL RIDS/DBS DBS/UTIL
    DEP: NONE
    GEN:
  
  *** END IMS KEYWORDS ***
  DRF early termination has a timing window where FRXQBUF0 is
  notified to continue processing even if an IO error has
  occurred that terminates DRF early.  Code is added to the
  log read component of DRF to ensure the timing window is
  closed.
  
  The following changes are made to modules in the log read
  component:
  
  FRXRCTL0:  Term early if notified to do so, even if there
  is a log still to read.  Only schedule another reader if
  there is no allocation errors.  Save the buffer address
  when clearing the RVRA when another reader is scheduled.
  
  FRXRDTH0:  Move the entry trace to when FRXRDTH0 is
  entered for diagnostics instead of in StartReading.
  After setting up the read and returning from StartReading
  subroutine, check the return code and set the continue
  flag if the return code is zero.
  If there is no buffer when entering the StartReading
  subroutine, get a buffer.  When an alloc or open error is
  detected and a copy of the log exists, turn off the
  LDSD_copy flag.
  Only trace the exit if there is no IO error.
  If there was an IO error, only notify FRXRCTL0 that we are
  done if we can not switch to a copy.
  
  FRXRBUF0:  If we have a read error, only tell FRXRCTL0 that
  read has finished early if there is no copy to switch to.
  

Temporary fix

Comments

APAR Information

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

  PK38717

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

  UK27281

Modules/Macros

• FRXRBUF0 FRXRCTL0 FRXRDTH0
  

Fix information

• Fixed component name

  IMS DB RECOVERY

• Fixed component ID

  5655I4400

Applicable component levels

• R310 PSY UK27281

     UP07/07/24 P F707