PH38307: CICS REGION USING NAMED COUNTER SERVER GETS ABEND 80A 878 AT SHUTDOWN WHEN TCBS ATTEMPT TO DISCONNECT FROM SHARED SERVER

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• Your CICS region using named counter server.  When shutting
  down the region, you get the following messages
  reporting 80A and 878 abends:
  
    IEA705I ERROR DURING GETMAIN SYS CODE = 80A-10
    IEA705I ERROR DURING GETMAIN SYS CODE = 878-10
  
  The last CICS message issued was:
    DFHRM0130 applid Recovery manager has successfully quiesced.
  
  In the dump, the DM domain shows:
        CICS had quiesced when this dump was taken
  
  Focusing on the storage usage below the line, the growth
  from the top down - LSQA/SWA/229/230 was high - over 1.75M.
  
  The failing getmains were for x'148', out of subpool 229,
  issued from shared server module AXMLFGET.
  
  Browsing some of these allocated pieces of storage, they
  had the eyecatcher DFHNC - for named counter server
  There were about 170 different TCB's attempting to disconnect
  from the server. Each TCB requests this 148 piece of storage,
  which actually allocates an entire page of storage - x'1000'
  bytes.
  Additional symptoms: KIXREVRJL
  AXMSC ENQ SYSZAXMC DFHNC
  

Local fix

• see if DSALIM could be lowered, to give more storage back to
  z/OS to satisfy getmains below the line.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS users.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: A CICS region abends 80A during         *
  *                      shutdown when TCBs disconnect from      *
  *                      the named counter server.               *
  ****************************************************************
  A CICS region had many L8 TCBs that had connected to a named
  counter server to process NCS requests for their programs.
  When CICS was shut down, these TCBs had to disconnect
  from AXM as part of the termination process. AXMLFMVS was called
  to acquire stack storage for the disconnect operation. The stack
  storage was for subpool 229, allocated from below the line. Each
  stack of x140 bytes required a page of subpool 229 storage if
  one was not already allocated. As many hundreds of L8 TCBs all
  required this storage at the same time, MVS happened to run out
  of available storage and the abend 80A resulted.
  

Problem conclusion

• AXMLFMVS has been changed to use GETMAIN RU and FREEMAIN RU to
  acquire and free above the line stack storage in this case.
  

Temporary fix

Comments

APAR Information

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

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

  UI77922 UI77923

Modules/Macros

• AXMLFMVS
  

Fix information

• Fixed component name

  CICS TS Z/OS V5

• Fixed component ID

  5655Y0400

Applicable component levels

• R200 PSY UI77923

     UP21/11/04 P F111

• R300 PSY UI77922

     UP21/11/04 P F111

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.