IBM Support

OA14653: POOR HPR PERFORMANCE WHEN RTP ARB DATA IS PROCESSED OUT-OF-SEQUENCE.

A fix is available

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • ARB Rate requests are being presented to RTP Out-of-sequence.
    RTP interpreted the data as erratic Delay Change Sum variations,
    which led to multiple Slowdowns and rate-of-flow adjustments.
    The changes caused unwarrented delays in the flow of a subset of
    the UDP/IP traffic over the pipe.  The RTP pipe flow eventually
    recovered, but not before the bottleneck was visible.
    

Local fix

  • A RTP Pathswitch will also reset the affected values for the
    Pipe.  Issue an appropriate "F NET,RTP,ID=pu-name" command.
    

Problem summary

  • ****************************************************************
    * USERS AFFECTED: All running with RTP/HPR connections.        *
    ****************************************************************
    * PROBLEM DESCRIPTION: Intermittent performance problems       *
    *                      experienced on RTP connection.          *
    ****************************************************************
    * RECOMMENDATION:                                              *
    ****************************************************************
    The problem is summarized as follows:
    
    Each end of an RTP pipe has sender and receiver logic.  This
    problem is in the receiver logic of a z/OS CS RTP
    endpoint.
    
    1) An RTP pipe is established and data traffic is
       occurring.
    2) VTAM is receiving ARB rate requests, as expected, from
       the RTP partner.
    3) At some point, two ARB rate requests (correlators F and 0)
       are within six milliseconds of each other on the TCP/IP host
       over EE).  SRBs are scheduled to process each inbound unit
       of work.
    4) Due to the nature of SRBs, it is possible for ARB (0)
       rate request to reach the RTP layer before the first one
       ARB (F).  This will cause RTP to process these ARB rate
       requests out of order.
    5) When the RTP layer processes ARB (0) rate request, it
       believes the ARB rate request (F) was lost in the network.
    6) The ARB (0) rate request was processed and an ARB rate reply
       returned to the partner.
    7) Now, the RTP layer receives the ARB (F) rate request out of
       order.  Invalid logic allowed the RTP layer to process the
       ARB rate request.  It was treated as if the next 14
       sequential ARB rate requests (correlators 1 through E) were
       lost in the network (when it is really out of order).
    8) This caused RTP to roll in an enormously large delay change
       (DC) into the delay change sum (DCS).  In this case, it was
       roughly 700 milliseconds.
    9) The invalid DCS causes VTAM to respond to subequent ARB
       rate requests with slowdowns.  These slowdowns indicate to
       the partner that it must cut its sending rate.  Since the
       DCS has an invalid value, VTAM repeatedly sends slowdowns
       to the ARB rate requests. Eventually, the RTP endpoints
       backup and endpoint devices and applications timeout as a
       result.
    

Problem conclusion

  • Module ISTRPCAR has been corrected to process the out of
    sequence ARB rate requests correctly.  If the received ARB rate
    request is equal to, or up to three less than, the last
    processed ARB rate request correlator, it is assumed to have
    arrived out of order and is ignored.  All other ARB rate
    requests are processed as usual.
    
    Module ISTRPCRT has been included for maintenance purposes.
    

Temporary fix

  • *********
    * HIPER *
    *********
    

Comments

APAR Information

  • APAR number

    OA14653

  • Reported component name

    VTAM V4 MVS/ESA

  • Reported component ID

    569511701

  • Reported release

    160

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    YesHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2005-12-15

  • Closed date

    2006-01-04

  • Last modified date

    2006-04-03

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

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

    UA23625 UA23626 UA23627 UA23628

Modules/Macros

  • ISTRPCAR ISTRPCRT
    

Fix information

  • Fixed component name

    VTAM V4 MVS/ESA

  • Fixed component ID

    569511701

Applicable component levels

  • R140 PSY UA23625

       UP06/03/02 P F603 Ž

  • R150 PSY UA23626

       UP06/03/02 P F603 Ž

  • R160 PSY UA23627

       UP06/03/02 P F603 Ž

  • R170 PSY UA23628

       UP06/03/02 P F603 Ž

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":"160","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":"160","Edition":"","Line of Business":{"code":"","label":""}}]

Document Information

Modified date:
03 April 2006