PK42909: RECOVERY HANGS AFTER MESSAGE FRD4214I

APAR status

• Closed as program error.

Error description

• DRF job goes into a wait state. DRF batch job was submitted to
  recover 4 HALDB partitions. Each partition had been split into 2
  parts, A & B, so  a total of 8 HALDB datasets needed to be
  recovered. The DRF job spun off 8 SASs, 7 of which completed, 1
  remained in a wait state. The 7 SASs  that completed all
  initialized (FRD1000I), requested a tape mount for
  the image copy, processed and then shutdown (BPE0009I). The SAS
  that went into a wait state initialized (FRD1000I) but never
  asked for a tape mount as it should have. Eventually the DRF
  batch job was cancelled, resulting in the following messages in
  the hung SAS region:
   BPE0006I BPE JSTPTCB ABEND S222
   FRD4214I RECORRD PIPE FAILURE DETECTED, WRITE, RC=C
   FRD2885I WRITE TERMINATION COMPLETE
   BPE0007I DRF BEGINNING PHASE 1 SHUTDOWN
                  TERMINATED AT END OF MEMORY
                  FAILED IN ADDRESS SPCE 0230 994
                  SYSTEM ABEND S069 REASON CODE 04
   The READNUM value was set to 5,20 and this seemed to work.
  There were 5 mounts issued from 5 of the 8 SASs, when one of
  these SASs completed a new SAS requested a tape mount, except
  for the one SAS that went into the wait/hang. There is a ASID
  dump of hung SAS, taken just prior to the cancel. This scenerio
  occurred 2 or 3 times one morning.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: Users with IMS Database Recovery Facility    *
  *                 Version 3 Release 1 installed.               *
  ****************************************************************
  * PROBLEM DESCRIPTION: DRF may run into a hang state when      *
  *                      attempting to restore from image        *
  *                      copies that reside on tape.  This       *
  *                      occurs in situations where the user     *
  *                      specified a READNUM value that is less  *
  *                      than the number of started subordinate  *
  *                      address spaces.                         *
  ****************************************************************
  * RECOMMENDATION: INSTALL CORRECTIVE SERVICE FOR APAR/PTF      *
  ****************************************************************
  DRF considers the value supplied by READNUM to be the number of
  tape devices available for processing image copies from tape.
  DRF will grant that number of subordinate address spaces to
  proceed and save their information on a "tape device ownership
  queue".  All other subordinate address spaces are placed on a
  "wait queue".  As each subordinate address space finish
  processing its image copy it is removed from the "tape device
  ownership queue".  The next subordinate address space is
  expected to be placed on the "tape device ownership queue" and
  removed from the "wait queue", however, this did not occur.
  

Problem conclusion

• AIDS: RIDS/UTIL RIDS/DBS DBS/UTIL
    DEP: NONE
    GEN:
  
  *** END IMS KEYWORDS ***
  FRXMSTR0 is changed to grant the next subordinate address space
  permission to process its image copy by including it on the
  "tape device ownership queue" and removing it from the "wait
  queue".
  

Temporary fix

Comments

APAR Information

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

  PK42203

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

  UK24124

Modules/Macros

• FRXMSTR0
  

Fix information

• Fixed component name

  IMS DB RECOVERY

• Fixed component ID

  5655I4400

Applicable component levels

• R310 PSY UK24124

     UP07/04/19 P F704