PM78250: CICS SHUTDOWN HANG

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• CICS should not be causing a shutdown hang because it is getting
  an
  error on ECI socket requests. We have made changes in the past
  but they
  did not cover what appears to be a new error so we would like to
  extend the range of problems that are handled.
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All CICS users with ECI TCPIPSERVICEs.       *
  ****************************************************************
  * PROBLEM DESCRIPTION: CICS shutdown hang after msgDFHWK0105I  *
  *                      when ECI TCPIPSERVICEs are being used   *
  *                      and SO0002 abends have occurred.        *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  Customers using ECI TCPIPSERVICEs may experience a CICS shutdown
  hang after SO0002 abends.
  
  The symptoms are identical to those described by APARs PM25851
  and PM44437. CICS will fail to shut down normally and hang after
  message DFHWK0105I WARM KEYPOINT SUCCESSFUL and a dump will
  show a task in a SODOMAIN SO_LTEPTY wait state. DFHPD410 SO=3
  will show one or more STEs in a closed state with the number of
  the task that failed to complete normally.
  
  If CICS trace is available for any SO0002 abend, it will show
  an exception entry for a socket receive with an errno of 157
  (EMVSERR):
  
  SO 0211 SOCK  *EXC* ASYNCIO_ERROR         1,157,000000FF
  
  This errno is normally a serious error but can be produced by
  TCP/IP to represent a failing client connection and possibly
  the socket being reused because the $SOCKOPT $OPTCHKR action
  has been taken to allow a dead socket to be reused by a
  different task. $OPTCHKR is an option of the CSI/IBM TCP/IP
  product not the BSI/IBM version.
  
  EMVSERR is reported to DFHIEIE as sock_IO_error, it assumes
  that there has been a serious error and it does not attempt to
  perform the full cleanup of the ECI client environment, which
  includes removing the STE control block from the LTE (the
  TCPIPSERVICE). During normal shutdown, CICS finds the queued
  STEs and waits for them to be deleted, which can never happen,
  and enters an indefinite wait state. CICS must be cancelled.
  
  Additional keywords: MSGDFHWK0105I WK0105I WK0105
                       DFHSO0002 msgDFHSO0002
  

Problem conclusion

• DFHIEIE has been changed so that a socket receive function will
  treat sock_IO_error as a connection failure and perform a full
  cleanup of the client and socket states.
  

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:

  UK91154

Modules/Macros

•    DFHIEIE
  

Fix information

• Fixed component name

  CICSTS FOR VSE

• Fixed component ID

  564805400

Applicable component levels

• RB0P PSY UK91154

     UP13/02/07 P E512

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.