IBM Support

PM71528: TCP TIMER POP WITH END_REF USES EXCESSIVE ECSA

A fix is available

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • The final timer expiration for a TCP connection (FINWAIT2 or
    TIMEWAIT) will cause the use count for the associated control
    blocks to drop to zero.  When this occurs, the EZBTCERF entry
    point in EZBTCUTL is invoked causing a DUSA expansion.  This
    will cause a 36K block of ECSA storage to be allocated for this
    processing.
    
    When multiple connections are terminated at the same time (such
    as the shutdown of a server application), the potential exists
    for all of these to have the final timer expire at the same.
    If there is resource contention involved in that final
    processing (such as communication for distributed connections or
    SMF recording), there can be a large number of these 36K areas
    generated.  This can lead to:
    
     - Various ABENDs, typically multiple S4C5/74FF2500 in TCPIP
       processing.
    
     - If ECSALIMIT is configured in TCPIP, EZZ4360I ECSA
       CONSTRAINED and EZZ4361I ECSA CRITICAL messages could be
       issued.
    
     - If SYSPLEXMONITOR is enabled in TCPIP, this situation can
       cause the stack to leave the sysplex.  This would be
       accompanied by EZD1973E MULTIPLE TCPIP NONRECOVERABLE ERRORS
       ARE ADVERSELY AFFECTING SYSPLEX PROCESSING and EZZ9676E
       SYSPLEX PROBLEM DETECTION CLEANUP HAS SUCCEEDED FOR TCPIP
       messages.
    

Local fix

  • If the stack has left the sysplex, use system monitoring tools
    to check the levels of ECSA usage and DISPLAY TCPIP,,STOR
    commands to monitor the amount of ECSA used by TCPIP.  Once the
    usages drop back to normal levels, a
    VARY TCPIP,,SYSPLEX,JOINGROUP can be issued to allow the stack
    to rejoin the sysplex.
    

Problem summary

  • ****************************************************************
    * USERS AFFECTED: All users of the IBM Communications Server   *
    *                 for z/OS Version 1 Release(s) 13 IP          *
    ****************************************************************
    * PROBLEM DESCRIPTION: Temporary ECSA storage growth occurs    *
    *                      when TCP connections are freed from     *
    *                      timer related processing.               *
    ****************************************************************
    * RECOMMENDATION:                                              *
    ****************************************************************
    Temporary ECSA storage growth occurs when TCP connections are
    freed from timer related processing.
    
    The TCP connection control block (TCB) contains a reference
    count.  When the reference count becomes zero a routine is
    called to perform cleanup processing on the TCB.  The module
    that most frequently performs the reference decrement to zero
    also contains the clean up routine.  The call to the clean up
    routine is always done as an external call which requires
    additional dynamic storage.  The clean up routine should be
    called as an internal routine to reduce the dynamic storage
    requirements.
    The TCP state timer routines for FINWAIT2 and TIMEWAIT use
    ECSA for dynamic storage.  Timer processing can schedule a
    large number of FINWAIT2 or TIMEWAIT timer routines that will
    most likely decrement the reference count to zero, thus causing
    a temporary increase in ECSA usage.  Converting the call to the
    TCB clean up routine to an internal call will reduce the ECSA
    usage when a large number of timer routines are scheduled at
    the same time.
    +-------------------------------------------------------------+
    + Please check our Communications Server for OS/390 homepages +
    + for common networking tips and fixes.  The URL for these    +
    + homepages can be found in Informational APAR II11334.       +
    +-------------------------------------------------------------+
    

Problem conclusion

  • EZBTCUTL has been amended to call the TCB clean up routine as
    an internal call.
    
    * Cross Reference between External and Internal Names
    

Temporary fix

Comments

  • ž**** PE12/11/14 FIX IN ERROR. SEE APAR PM77118  FOR DESCRIPTION
    

APAR Information

  • APAR number

    PM71528

  • Reported component name

    TCP/IP V3 MVS

  • Reported component ID

    5655HAL00

  • Reported release

    1D0

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2012-08-24

  • Closed date

    2012-09-20

  • Last modified date

    2012-12-04

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

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

    UK82041 PM76819

Modules/Macros

  • EZBTCUTL TOTCPDS
    

Fix information

  • Fixed component name

    TCP/IP V3 MVS

  • Fixed component ID

    5655HAL00

Applicable component levels

  • R1D0 PSY UK82041

       UP12/10/31 P F210

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":"BU054","label":"Systems w\/TPS"},"Product":{"code":"SG19M","label":"APARs - z\/OS environment"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"1D0","Edition":"","Line of Business":{"code":"","label":""}},{"Business Unit":{"code":"BU054","label":"Systems w\/TPS"},"Product":{"code":"SSCY4DZ","label":"DO NOT USE"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"1D0","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
04 December 2012