IBM Support

PI62074: TCPIP TIMERS MAY NOT BE EXPIRING AT CORRECT INTERVALS RESULTING IN UNEXPECTED DELAYS IN TIMER-BASED ACTIVITIES

A fix is available

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • With TCPIP timers not expiring at the correct intervals, this
    can lead to potential delays in any of the timer-based
    activities such as TCP packet retransmissions, TCP keepalive
    packets, TCP delayed acknowledgements, Sysplex monitoring, etc.
    
    Possible symptoms seen:
    
    * Sysplex monitor timer
    
      EZZ9674E tcpip_proc SYSPLEX PROCESSING WAS NOT RESPONSIVE FOR
              AT LEAST timer_secs SECONDS
    
      If RECOVERY was specified on the SYSPLEXMONITOR parameter of
      the GLOBALCONFIG statement, and this TCPIP stack is not the
      only member of the TCPIP sysplex group, the TCPIP stack leaves
      the sysplex group, inactivates all DVIPAs, saves all of
      VIPADYNAMIC block definitions, and takes S4C5 abend dump:
    
      EZZ4215I TCP/IP ABEND - DUMPING
      IEA794I SVC DUMP HAS CAPTURED:
        DUMPID=xxx REQUESTED BY JOB (tcpip_proc)
        DUMP TITLE=tcpip_proc  VxRx: Job(tcpip_proc)
        EZBXFPDX(HIP6xxx mm.ddd)+00045A S4C5/77620405
        SRBP=tcpip_asid,S=tcpip_asid,H=tcpip_asid
      EZZ9676E SYSPLEX PROBLEM DETECTION CLEANUP HAS SUCCEEDED FOR
               tcpip_proc
    
    Dump analysis:
    
    * Sysplex monitor timer problem determination events
    
      EZBXFCVT: 1F0D0000
      +06BE  XFCVT_PD_EVENTS................. 01000000
    
      Only bit7 in byte0 x'01' flag is ON for XFCVT_PDCSUCC_EVENT
      for successful event and all other events (OMPROUTE, VTAM,
      CSM, XCF Routes, Sysplex abend, and XCF status update) are
      OFF to indicate no error events. That is:
    
      0123 4567
      1... ....   XFCVT_OMP_EVENT1      initial OMP unrsp msg
      .1.. ....   XFCVT_OMP_EVENT2      last OMP unrsp msg
      ..1. ....   XFCVT_VTAM_EVENT      VTAM unresponsive msg
      ...1 ....   XFCVT_CSM_EVENT       CSM critical msg
      .... 1...   XFCVT_XCFROUTE_EVENT  No XCF routes msg
      .... .1..   XFCVT_Retry_EVENT     Sysplex Abend msg
      .... ..1.   XFCVT_XCFSTAT_EVENT   XCF Status msg
      .... ...1   XFCVT_PDCSUCC_Event   PDC success event act
    
    * Sysplex monitor timer timestamps for the interval updates:
    
      +0650  XFCVT_VTAMDOWN_TIME............. 00000000  00000000
      +0658  XFCVT_OMPROUTEUP_TIME........... D0AECDFE  93EA5F86
      +0660  XFCVT_XCF_ROUTES_TIME........... D0AECDBB  9265DE85
      +0668  XFCVT_STAT_TIME................. D0AECDDC  B9F52100
      +0670  XFCVT_MONINTF_TIME.............. D0AECDBB  9265DE85
      +0678  XFCVT_DYNROUTE_TIME............. D0AECDBB  9265DE85
    
      VTAM Down Time     00:00:00.000000
      OMPROUTE Up Time   22:13:02.403237 <- OMPROUTE is active
      XCF Routes Time    22:11:52.142429 <- Last update 70 secs ago
      Status Update Time 22:12:26.907474 <- Last update 35 secs ago
      MonINTF Time       22:11:52.142429 <- Last update 70 secs ago
      DynROUTE Time      22:11:52.142429 <- Last update 70 secs ago
    
      22:13:02.57 EZZ9674E tcpip_proc SYSPLEX PROCESSING WAS NOT
                  RESPONSIVE FOR AT LEAST 66 SECONDS
    
      The XCF status update timer interval should be 1/4 of the
      TIMERSECS value. In this case with "TIMERSECS 66" coded,
      the interval is roughly 16 seconds. If no timer updates were
      made to the XCF status at the intervals such that the delay
      has exceeded the TIMERSECS value, then EZZ9674E gets issued.
    
      The Status Update Time is when EZBXFPDX was scheduled the
      first time because EZBXFPDM had not run and updated the time
      stamp. The XCF Routes Time is how long ago it was when PDM
      last ran.
    

Local fix

  • * For sysplex monitor timer problem:
    
      1) Specify NORECOVERY on the SYSPLEXMONITOR statement so that
         no further recovery actions are taken. That is, the TCPIP
         stack will remain in the sysplex group. When using this
         option, ensure that the TCPIP stack is healthy based on no
         error conditions from the NETSTAT and storage reports.
    
      2) If RECOVERY is specified on the SYSPLEXMONITOR statement,
         set up automation to monitor for the EZZ9674E/EZZ9676E
         messages and issue the VARY TCPIP,,SYSPLEX,JOINGROUP
         command to rejoin the TCPIP stack to the sysplex group.
         Ensure that the TCPIP stack is healthy based on no error
         conditions from the NETSTAT and storage reports.
    
    KEYWORDS:
    TCP TIMERS ABEND S4C5 EZBITTUB EZBXFPDX EZBXFPDM EZZ9674E
    SYSPLEX MONITOR XMIT RETRANSMIT KEEPALIVE DELAYACK ACK
    TIMERSECS
    

Problem summary

  • ****************************************************************
    * USERS AFFECTED:                                              *
    * All users of the IBM Communications Server                   *
    * for z/OS Version 2 Releases 1 and 2 IP                       *
    ****************************************************************
    * PROBLEM DESCRIPTION:                                         *
    * TCPIP timers expire past the specified interval              *
    ****************************************************************
    * RECOMMENDATION:                                              *
    * Apply PTF                                                    *
    ****************************************************************
    The TCPIP timer set processing may incorrectly calculate the
    expiration time to a greater value than the requested delta.
    The calculation error only occurs when a timer is being set at
    the same time as the TCPIP clock management task is executing
    and the internal clock management table wraps.
    

Problem conclusion

  • The timer set processing has been amended to take into account
    
    the wrapping of the internal clock management table.
    

Temporary fix

Comments

APAR Information

  • APAR number

    PI62074

  • Reported component name

    TCP/IP V3 MVS

  • Reported component ID

    5655HAL00

  • Reported release

    210

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    YesHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2016-05-06

  • Closed date

    2016-06-03

  • Last modified date

    2016-07-26

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

    PI62051

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

    UI38314 UI38315

Modules/Macros

  • EZB2FCVT EZBXFINI EZBITTUB EZBXFPDM EZBTIGBL
    

Fix information

  • Fixed component name

    TCP/IP V3 MVS

  • Fixed component ID

    5655HAL00

Applicable component levels

  • R210 PSY UI38314

       UP16/06/29 P F606

  • R220 PSY UI38315

       UP16/06/29 P F606

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":"210","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":"210","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
26 July 2016