IBM Support

PH33864: NEW REORG OPTION TO MINIMIZE DURATION OF LAST LOG ITERATION

A fix is available

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as new function.

Error description

  • New REORG option to minimize duration of last log iteration
    

Local fix

Problem summary

  • ****************************************************************
    * USERS AFFECTED:                                              *
    * All Db2 12 for z/OS users of REORG                           *
    * SHRLEVEL CHANGE utility                                      *
    ****************************************************************
    * PROBLEM DESCRIPTION:                                         *
    * New function for REORG SHRLEVEL CHANGE                       *
    * utility to minimize duration of the                          *
    * last log iteration in the LOG phase                          *
    ****************************************************************
    * RECOMMENDATION:                                              *
    ****************************************************************
    This APAR introduces a new keyword option to the REORG SHRLEVEL
    CHANGE utility, LASTLOG YES|NO, to further minimize the
    application outage window in the last log iteration in the LOG
    phase.
    During the LOG phase of the REORG SHRLEVEL CHANGE utility, the
    utility would quiesce all of the read and write claimers, then
    perform one last round of log apply before advancing to the
    SWITCH phase.  Depending if there are any log records applied
    to the shadow data sets during this final log iteration, REORG
    might need to perform actions such as force write and final
    incremental copy to reflect the latest DML changes.  In order
    to minimize these processing costs during the outage duration,
    a new parameter LASTLOG YES|NO is added to the REORG utility,
    as described below.
    For REORG TABLESPACE utility syntax diagram:
    >--REORG TABLESPACE-->
    Change-spec:
          |--DELAY 1200-----|  |--LASTLOG YES--|
    >-----|-----------------|--|---------------|--->
          |--DELAY integer--|  |--LASTLOG NO---|
    For REORG INDEX utility syntax diagram:
    >--REORG--|--INDEX LIST listdefname--|--->
              |--index-name spec---------|
    Change-spec:
        |--DELAY 1200-----| |--LOGRANGES YES--| |--LASTLOG YES--|
    >---|-----------------|-|-----------------|-|---------------|->
        |--DELAY integer--| |--LOGRANGES NO---| |--LASTLOG NO---|
    LASTLOG
    Specifies whether the REORG SHRLEVEL CHANGE utility would apply
    any log records in the final log iteration in the LOG phase.
    By the specification of LASTLOG NO, it can help reduce the
    outage window by avoiding the costly sub processes (such as
    page sets force write) that occur in the final log iteration,
    but it might cause REORG to not be able to complete if there
    is no quiet window with the absence of concurrent DML
    activities.  LASTLOG specification is ignored for non-SHRLEVEL
    CHANGE processing, and LASTLOG NO requires the DRAIN ALL option
    to be in effect.
    YES
    REORG utility is to perform one final round of log apply after
    acquiring the drain all.  This ensures that the REORG would
    proceed to the SWITCH phase after completing the final round of
    log apply in the LOG phase. The default value is YES.
    NO
    REORG utility is to ensure that no log records are to be
    applied in the final log iteration.  When existing criteria is
    met for REORG to break in, REORG would first perform the drain
    all and then process the logs from the end of the last log
    iteration to current.  If any log records of the target objects
    are found in this final log iteration, REORG would dedrain the
    target objects and revert this final log iteration back to a
    normal log iteration.  When it is time to break in again in a
    future log iteration, REORG would repeat this cycle of drain
    all, log read, and dedrain logic until it can complete the
    final log iteration with no log records to apply. The reversion
    back to a normal log iteration due to the presence of logs
    counts as a drain failure for RETRY consideration, so a high
    RETRY value is recommended to lessen the impact of the
    repeating break in attempts.
    

Problem conclusion

Temporary fix

Comments

  • Code has been modified to introduce the aforementioned
    new function to the REORG utility.
    

APAR Information

  • APAR number

    PH33864

  • Reported component name

    DB2 OS/390 & Z/

  • Reported component ID

    5740XYR00

  • Reported release

    C10

  • Status

    CLOSED UR1

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt / Xsystem

  • Submitted date

    2021-01-27

  • Closed date

    2022-04-19

  • Last modified date

    2022-05-17

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

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

    UI80216 PH46524

Modules/Macros

  • DSNUGDDM DSNURLOG DSNUGUCA DSNUGPRT
    

Fix information

  • Fixed component name

    DB2 OS/390 & Z/

  • Fixed component ID

    5740XYR00

Applicable component levels

  • RC10 PSY UI80216

       UP22/04/28 P F204

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.

[{"Line of Business":{"code":"LOB10","label":"Data and AI"},"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSEPEK","label":"DB2 for z\/OS"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"12.0"}]

Document Information

Modified date:
18 May 2022