IBM Support

PK99036: DFHSO0001 ABEND ASOB AT X'0A96' IN DFHSOS14. TI TASK RUNNING ON SO TCB INSTEAD OF QR

A fix is available

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • In module DFHSOS14 (level UK46790) +A96 is where CICS attempts
    to free a lock, and the call to LMLM did not return with
    lmlm_ok.
    .
    In the registers from the Kernel Error entry, R8+x'1A' contains
    the actual lmlm_response. There we see x'02', which equates to
    LMLM_EXCEPTION.  R8 is the address of the DFHLMLM_ARG, the
    parameters. It is x'48' bytes long. The reason for the exception
    is in LMLM_REASON = LMLM_LOCK_TOKEN_NOT_FOUND.
    .
    If trace is turned on, leading up to the failure you see the TI
    task wake up to cancel an ASYNC receive for a socket. It starts
    to switch to the SO TCB. However, before the task can be
    dispatched on the SO TCB the sockets listener CSOL task wakes up
    because the ASYNC RECEIVE has completed because the client has
    closed the connection from its end.
    .
    CSOL switches to the SO TCB too. CSOL cancels the ASYNC RECEIVE,
    closes the socket from the CICS end and tries to freemain an
    ASYNCBUF addressed by the CICS socket object. This loses control
    on the SMLOCK.
    .
    Now the TI task gets dispatched on the SO TCB. The TI task tries
    to cancel the RECEIVE which fails (because CSOL has already
    canceled it). The TI task completes soso_cancel_inbound and does
    not switch off of the SO TCB at this point.
    .
    Additional Keywords:
    DFHSO0001 An abend (code 0C4/AKEX) has occurred at offset
    X'05EC' in module DFHSOS09.
    
    KIXREVRBD
    

Local fix

Problem summary

  • ****************************************************************
    * USERS AFFECTED: All                                          *
    ****************************************************************
    * PROBLEM DESCRIPTION: MSGDFHSO0001 ABENDASOB caused by long   *
    *                      running TI task executing on SO TCB     *
    *                      instead of QR TCB.                      *
    ****************************************************************
    * RECOMMENDATION:                                              *
    ****************************************************************
    The long running TI task was attempting to timeout a socket
    RECEIVE due to the SOCKETCLOSE timeout time being hit.
    In parallel, a web task was attempting to relinquish the
    reserve it had on the socket during task termination.
    Timeout processing caused the socket to be closed and deleted
    just as the web task was in relinquish processing. This
    resulted in an ABENDASOB due to a lock problem within
    relinquish logic.
    Normally, task termination socket relinquish processing and
    SOCKETCLOSE timeout processing would be single threaded as both
    processes should run on the QR TCB.
    However, the long running TI task had got stuck on the SO TCB
    due to an earlier problem. This problem occurred when processing
    a SOCKETCLOSE TIMEOUT for an earlier socket. In parallel with
    the SOCKETCLOSE TIMEOUT on the TI task, the sockets listener
    task ( CSOL ) was attempting to close the same socket because
    the client had closed it.
    This resulted in a socket synchronization issue between the TI
    task and CSOL which caused the TI task to get stuck on the
    SO TCB.
    
    ADDITIONAL KEYWORDS :- SO0001
    

Problem conclusion

  • DFHSOTI (the module which handles SOCKETCLOSE timeout) has been
    changed to ensure that it exits on the same TCB mode on which
    it enters.
    

Temporary fix

  • FIX AVAILABLE BY PTF ONLY
    

Comments

APAR Information

  • APAR number

    PK99036

  • Reported component name

    CICSTS V3 Z/OS

  • Reported component ID

    5655M1500

  • Reported release

    500

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2009-10-20

  • Closed date

    2010-01-07

  • Last modified date

    2010-02-01

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

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

    PM02806 UK53435

Modules/Macros

  •    DESSOTI  DFHSOTI
    

Fix information

  • Fixed component name

    CICSTS V3 Z/OS

  • Fixed component ID

    5655M1500

Applicable component levels

  • R500 PSY UK53435

       UP10/01/12 P F001

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":"3.2","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":"3.2","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
01 February 2010