PK24822: ACTSSLTCBS ALWAYS AT 1 EVEN THOUGH MAXSSLTCBS IS 8 AND LOTS OF TASKS ARE STUCK IN DISPATCH SSL_POOL WAITING ON AN S8 TCB .

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• There is only 1 S8 TCB available for use in the SSL POOL of
  TCBs even though MAXSSLTCBS is set to 8 and lots of tasks are
  waiting in DISPATCH SSL_POOL waits waiting for a S8 TCB to run
  on.
     New S8 TCBs are not attached because the current number of
  TCBs attached in the SSL Pool is at 8, the maximum.  Yet the
  current number of TCBs attached in the S8 Mode is 1.
     Here is what this phenomenon would look like externally:
  A CEMT INQ DISPATCHER would show
  Actssltcbs(0001)
  Maxssltcbs( 0008 )
  .
  A CEMT INQ TASK would show lots of tasks with
  Htype(DISPATCH)
  Hvalue(SSL_POOL)
  .
  A DFH0STAT Dispatcher Pool report shows
  TCB Pool . . . . . . . . . . . . . . . . . :             SSL
  Current TCBs attached in this TCB Pool . . :               8
  Peak TCBs attached in this TCB Pool. . . . :               8
  Max TCB Pool Limit (MAXSSLTCBS). . . . . . :               8
  ...
  and
  ...
  Current TCBs in use in this TCB Pool . . . . . :           1
  Peak TCBs in use in this TCB Pool. . . . . . . :           1
  ...
  and
  ...
  TCB    TCBs Attached    < TCBs In Use ->
  Mode  Current    Peak   Current    Peak
  __________________________________________
   S8         1        1        1        1
  --------------------------------------------------------
  In a dump, you will see that there are 8 KTCB control blocks
  and 8 DS_TCB control blocks for 8 S8 TCBs.  Those 8 KTCBs at
  +X'50' contain MVS TCB addresses but only 1 of those actually
  exists as you can see when you use SUMM FORMAT to format the
  address space's TCBs.
  .
  7 of the DS_TCBs will have bits DELETE_TCB_REQUESTED (X'10' bit
  at +X'4D') and DELETE_TCB_COMPLETE (X'40' bit at +X'4D') set.
  .
  The OPEN_POOL control block for the SSL POOL
  ( DSANC.SSL_POOL xxxxxxxx OPEN TCB POOL ) has
  at +X'28' 00000008 (for CURR_OPEN_TCBS ) and at +X'3C' 00000007
  (for IN_TERM_NUM which is the number of TCBs in termination.
  .
  These S8 TCBs were deleted hours ago during a nightime period
  when the tasks that use the S8 TCBs stopped for the night.
  CICS then gradually all but one of the S8 TCBs.  But the
  process of deleting and detaching the TCBs has stopped in
  the middle.  7 of the 8 are stuck in termination.
  .
  The problem is that the post_detach code in DFHDSTI never runs
  for these TCBs.  This is where the CURR_OPEN_TCBS and
  IN_TERM_NUM fields would be decremented.  In prior releases,
  this code would normally have been driven by the QR TCB going
  through EVERY_SO_OFTEN_ACTIONS and making the DFHKEDSI
  DETACH_OWN_TERMINATED_TCBS call.  But at CICS/TS 3.1 the QR
  TCB is not the one that attaches the S8 TCBs.  In CICS/TS 3.1
  a new TCB attaches the S8s and that new TCB never invokes
  this code that would complete the detach.
  additional symptoms:
  DFHSO0002 A severe error (code X'0814') has occurred in
            DFHSOSE
  code 0814 after implementing SSL
  

Local fix

• Use CEMT SET DISPATCHER MAXSSLTCBS to increase the number of
   SSL TCBs.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All.                                         *
  ****************************************************************
  * PROBLEM DESCRIPTION: Lots of tasks stuck in DISPATCH         *
  *                      SSL_POOL waiting on S8 TCBs because     *
  *                      only 1 of 8 S8 TCBs are active.         *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  Over time inactive S8 TCBs will experience a DELETE_TCB which
  can never complete. They will be left in a 'being terminated'
  state. However while in this state they will still contribute
  to the MAXSSLTCBs limit.
  

Problem conclusion

• DFHDSAT has been altered so that we do not attempt to
  terminate inactive S8 TCBs.
  

Temporary fix

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

Comments

APAR Information

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

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

  UK18062

Modules/Macros

•    DFHDSAT
  

Fix information

• Fixed component name

  CICSTS 3.1 Z/OS

• Fixed component ID

  5655M1500

Applicable component levels

• R400 PSY UK18062

     UP06/09/20 P F609

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.