PK15899: JVM HUNG AFTER ABEND WHEN RUNNING WITH ONLY ONE JVM

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• 5697E9300
  Customer as MAXJVMTCBS=1 coded in DFHSIT. All is well until an
  LE ABEND occurs during the delete_tcb process of a Jx TCB (J9
  TCB in this case) the transaction was running on. A transaction
  dump was issued to report the abend. This causes a change mode
  back to the QR TCB to write to the dump datasets. Upon
  completion of the dump, a change mode back to the original J9
  TCB was attempted (same J9 TCB trying to be deleted). Dispatcher
  can not assign a new J9 TCB due to pool limit of 1 already being
  reached.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All                                          *
  ****************************************************************
  * PROBLEM DESCRIPTION: A CICS Java task hangs after a          *
  *                      transaction abend.                      *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  A CICS task running on a J8 or J9 TCB may hang following a
  transaction abend if a transaction dump is written and the
  JVM pool is at its limit (MAXJVMTCBS).
  
  During dump processing, a switch to the QR TCB is made to write
  to the transaction dump dataset. After the write is completed,
  a change back to the original TCB mode is requested. However,
  if the failing TCB is marked for deletion and the JVM pool is
  at its limit, the change back to the J8 or J9 TCB will fail.
  The failing task will hang with a suspend type of DISPATCH and
  suspend value of JVM_POOL until a TCB does become available.
  

Problem conclusion

• DFHDUIO has been changed to save the token for the original TCB
  before switching to the QR TCB for the dump i/o operation. The
  saved TCB token (rather than the mode) is then passed on the
  call to DFHDSAT to change mode.
  

Temporary fix

•             *********
              * HIPER *
              *********
  FIX AVAILABLE BY PTF ONLY
  

Comments

APAR Information

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

  PK11001

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

  UK10045

Modules/Macros

•    DFHDUIO
  

Fix information

• Fixed component name

  CICSTS 3.1 Z/OS

• Fixed component ID

  5655M1500

Applicable component levels

• R400 PSY UK10045

     UP05/12/17 P F512

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.