IBM Support

LI73140: REORG FAILURE FOLLOWED BY DROPPING OF TEMP TABLESAPCE CAN CAUSE CRASH RECOVERY TRAP DUE TO REORG LOG RECORD FALLING IN CR WINDOW

Subscribe

You can track all active APARs for this component.

 

APAR status

  • Closed as program error.

Error description

  • Scenario
    
    Temp tablespace created.  One or more reorgs are started using
    this temp tablespace and then subsequently interrupted.  Temp
    tablespace being used by reorgs is dropped.  Operating System
    shutdown/crash occurs which results in DB2 being abnormally
    shutdown.  Restart system and DB2.  Connect to database or
    activate database results in crash recovery . The reorg log
    records are within the recovery window and are replayed.
    References to the now deleted tablespace result in internals of
    DB2 using invalid pointers and a crash occurs during recovery.
    
    
    Avoidance
    
    In order to avoid the recovery crash it is necessary to remove
    the reorg log records from the recovery window.  In order to do
    this, the minBuffLSN (oldest committed transaction not yet
    hardened to disk and is the starting point for recovery) must be
    set past the last reorg log record Log Sequence Number.  The Log
    File Header itself then needs to be hardened to disk .
    
    To accomplish this , the following steps can be peformed :
    
    connect to db
    db2 create tablespace temp
    start reorgs
    reorgs are interrupted
    db2 force applications all (this will force all connection to be
    torn down at the server and force rollback of all inflight
    transactions)
    db2 deactivate db <dbname> (this will force log file closure and
    flush the log file header to disk).
    db2 connect to db
    db2 drop tablespace temp
    
    If a DB2 crash occurs at this point then reorg log records will
    not be within the recovery window and would therefore not be
    replayed and therefore the recovery crash would be avoided.
    

Local fix

  • Once the problem is hit no workaround. However, the situation
    could be avoided as described in this APAR
    

Problem summary

  • Users affected: User intrrupts reorg and then stops the
    instance abruptly
    
    Problem Description: Please  see APAR  ERROR DESCRIPTION
    
    Problem Summary: Interruption of reorg followed by dropping
    of temp tablespace can cause crash recovery failure after
    an abrupt instance failure
    

Problem conclusion

  • First fixed in DB2 UDB Version V8, FixPak 17
    (CMVC defect#379406)
    

Temporary fix

Comments

APAR Information

  • APAR number

    LI73140

  • Reported component name

    DB2 UDE ESE LIN

  • Reported component ID

    5765F4104

  • Reported release

    820

  • Status

    CLOSED PER

  • PE

    NoPE

  • HIPER

    NoHIPER

  • Special Attention

    NoSpecatt

  • Submitted date

    2008-02-05

  • Closed date

    2008-11-05

  • Last modified date

    2008-11-05

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

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

    LI73141 LI73142

Fix information

  • Fixed component name

    DB2 UDE ESE LIN

  • Fixed component ID

    5765F4104

Applicable component levels

  • R820 PSY

       UP

[{"Business Unit":{"code":"BU058","label":"IBM Infrastructure w\/TPS"},"Product":{"code":"SSEPGG","label":"Db2 for Linux, UNIX and Windows"},"Component":"","ARM Category":[],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"820","Edition":"","Line of Business":{"code":"LOB10","label":"Data and AI"}}]

Document Information

Modified date:
05 November 2008