IBM Support

PM15176: CICS SA03 ABEND. CIST TASK IN LMQUEUE WAIT. CSOL OWNS GLOBAL LOCK. IPIC SSL CONNECTIONS IN USE

A fix is available

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • Your CICS region abends with SA03 during shutdown and you find a
    hang involving CSOL owning a global lock but is suspended, while
    CIST wants the lock.   Dispatcher Domain for these tasks will
    look similar to the following
    .
    RESOURCE RESOURCE W TIME OF      TIMEOUT
    TYPE     NAME       SUSPEND      DUE
             LMQUEUE  S 00:01:45.105      -   <-- CIST
    SOCKET   SOCLOSE  S 00:01:45.105      -   <-- CSOL
    .
    The task that was running at the time that experienced the error
    that generated the SA03 dump was the Domain Manager task. It's
    KE stack will look similar to the following:
    .
    KE_NUM @STACK   LEN  TYPE ADDRESS  OFFSET NAME
    007C   4AE69020 0120 Bot  B9201C00 0003BC DFHKETA
    007C   4AE69140 0320 Dom  B921A640 000218 DFHDSKE
    007C   4AE69460 0460 Dom  B9275948 000AD4 DFHDMDS
                         Int   +000ADE 0001B2 SELECT_INIT_OR_QUIESCE
                         Int   +0008A8 000B5A QUIESCE
                         Int   +000A84 000904 QUIESCE_DOMAIN
    007C   4AE698C0 0D40 Dom  BA21F290 002CF4 DFHSODM
                         Int   +002EB4 0001AA QUIESCE_DOMAIN
                         Int   +002C6E 003050 DELETE_SSL_SUBTASKS
    007C   4AE6A600 09B0 Dom  B927B008 001A52 DFHDSIT
                         Int   +000542 0001B2 PROCESS
                         Int   +0018B8 0005A6 DELETE_TCB
    007C   4AE6AFB0 02D0 Dom  BA253E90 00027A DFHSODS
                         Int   +0001E4 00018E NOTIFY_DELETE_TCB
    007C   4AE6B280 03F0 Dom  BA8D1300 000000 DFHLEPT
                         Int   +00043C 0001C8 TERMINATE_LE_ENCLAVE
    .
    The problem is a caused by a timing window, and is only exposed
    if an IPIC connection is using SSL.
    .
    Additional Symptom(s) Search Keyword(s):
    KIXREVCEW
    

Local fix

Problem summary

  • ****************************************************************
    * USERS AFFECTED: All                                          *
    ****************************************************************
    * PROBLEM DESCRIPTION: Shutdown hang with a CIST task waiting  *
    *                      for the SOMGLOCK and CSOL waiting on    *
    *                      SOCLOSE suspend.                        *
    ****************************************************************
    * RECOMMENDATION:                                              *
    ****************************************************************
    A CICS system with an active IPCONN is shutdown. TCPIP activity
    is quiesced and a CIST task is attached by the sockets listener
    task as an IPIC TCPIPSERVICE is being closed. CIST issues a
    close of an outbound IPIC socket. This socket uses SSL so the
    close causes an alert to be sent to the remote SSL partner. This
    drives socket send processing. While this send is in progress
    CSOL completes the closure of the inbound sockets for the
    TCPIPSERVICE. There are no more active TCPIPSERVICEs so CSOL
    drives terminate_all to remove any remaining sockets.
    terminate_all obtains the global SOMGLOCK before looping through
    the unowned_sockets chain. The outbound socket which CIST is
    trying to close is on this chain. CSOL initiates close
    processing for this socket but detects that there is a socket
    SEND in progress. CSOL therefore suspends on resource name
    SOCLOSE. CSOL expects to be resumed by CIST when the socket send
    completes.
    When CIST completes its socket send io_completed is driven.
    io_completed contains code to resume CSOL. However, before it
    reaches this code it drives the setunowned function to move the
    socket to the unowned socket chain (even though the socket is
    already on the unowned chain).
    Setunowned tries to obtain the global SOMGLOCK. This is held by
    CSOL so CIST now suspends resulting in a deadlock.
    

Problem conclusion

  • The socket domain setunowned function has been changed to bypass
    the request to obtain the SOMGLOCK if the socket is already on
    the unowned chain.
    

Temporary fix

  • FIX AVAILABLE BY PTF ONLY
    

Comments

APAR Information

  • APAR number

    PM15176

  • Reported component name

    CICS TS Z/OS V4

  • Reported component ID

    5655S9700

  • Reported release

    600

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2010-05-25

  • Closed date

    2010-05-28

  • Last modified date

    2010-07-02

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

    PM12215

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

    UK57407

Modules/Macros

  • DESSOMG  DESSOSO  DFHSOMGC DFHSOM01 DFHSOM02
    DFHSOM03 DFHSOSOC DFHSOS00 DFHSOS01 DFHSOS02 DFHSOS03 DFHSOS04
    DFHSOS05 DFHSOS06 DFHSOS07 DFHSOS08 DFHSOS09 DFHSOS10 DFHSOS11
    DFHSOS12 DFHSOS13 DFHSOS14 DFHSOS15 DFHSOS16 DFHSOS17 DFHSOS18
    DFHSOS19 DFHSOS20 DFHSOS21 DFHSOS22 DFHSOS23
    

Fix information

  • Fixed component name

    CICS TS Z/OS V4

  • Fixed component ID

    5655S9700

Applicable component levels

  • R600 PSY UK57407

       UP10/06/04 P F006

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.

[{"Business Unit":{"code":"BU058","label":"IBM Infrastructure w\/TPS"},"Product":{"code":"SSGMGV","label":"CICS Transaction Server"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"4.1","Edition":"","Line of Business":{"code":"LOB35","label":"Mainframe SW"}},{"Business Unit":{"code":"BU054","label":"Systems w\/TPS"},"Product":{"code":"SG19M","label":"APARs - z\/OS environment"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"4.1","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
02 July 2010