PK39051: CICS ADDRESS SPACE HANGS WHEN CEJR IS EXCUTING A CALL TO THE DNS SERVER AND CICS IS SHUTDOWN

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• Customer is running transaction CEJR to perform a
  VERIFY_IP_ADDRESS which makes a call to DNS server. While
  executing the call to the DNS server CICS is shutdown. The
  shutdown hangs. The CICS regions receives a A03 abend followed
  by a KERNDUMP ( taken to address the abend 999 ) and eventually
  the regions receives a DFHSO0002 (X'0418') some 15 minutes after
  the shutdown was performed. The CICS address space never goes
  away.
  

Local fix

• If you experience this problem, we recommend you lower your
  TCP/IP RESOLVERTIMEOUT value from the default of 30
  seconds.
  

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS Users.                              *
  ****************************************************************
  * PROBLEM DESCRIPTION: Hang in CICS shutdown after CEJR        *
  *                      transaction was unable to resolve       *
  *                      an IP address.                          *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  A CORBASERVER is installed and CICS attaches transaction CEJR
  to resolve it. CEJR performs a DFHSOIS VERIFY_IP_ADDRESS which
  makes a call to a DNS server. This results in a CEEPIPI
  call to invoke TCPIP fuction gethostbyname to the DNS server.
  At this point the task is running on the SO TCB. As the DNS
  server does not respond immediately the Verify request waits
  until the TCPIP RESOLVERTIMEOUT value has elapsed.
  When it does, an error response of 'TRY_AGAIN' is returned
  and the Verify request is re-issued. This loop can be repeated
  up to 50 times.
  The RESOLVERTIMEOUT value was set to the default of 30 seconds
  so the loop resulted in a long wait. For the duration of this
  loop the SO TCB is unavailable to other tasks.
  All the domains shutdown but the CEJR task is still running
  on SO. Evetually CEJR gives up on its attempt to resolve
  the IP address. DFHKETCB tries to terminate the TCB but
  finds that there is a child TCB to quiesce first. It then
  waits for the child TCB to quiesce but this has already been
  deleted, so shutdown hangs.
  Additional keywords: HUNG SA03 abendA03 A03 abend SA03
  

Problem conclusion

• DFHSOIS has been altered to only attempt the gethostbyname
  call 5 times. Also, if the call returns a TRY_AGAIN response
  the routine will perform a DFHDSATM CHANGE_PRIORITY call to
  give up control and allow other tasks to run on the SO TCB.
  

Temporary fix

• 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:

  UK36602 UK36603

Modules/Macros

•    DESSOIS  DFHSOIS
  

Fix information

• Fixed component name

  CICSTS 3.1 Z/OS

• Fixed component ID

  5655M1500

Applicable component levels

• R400 PSY UK36602

     UP08/05/27 P F805

• R500 PSY UK36603

     UP08/05/27 P F805

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.