DB2 FOR LUW / DB2FORLUW / 970 - DB2

There is a very small timing hole that could result in an index 
scan not returning all of the qualifying rows.  The error can 
occur in one of two scenarios: 
 
1.  To hit the problem in scenario #1 all of the following must 
be true: 
- the index is defined using the MINPCTUSED keyword 
- a DELETE statement is executed such that the delete: 
  * has exclusive access to the table either obtained via the 
LOCK TABLE <tableName> IN EXCLUSIVE MODE statement or via 
lock escalation 
  * merges the right-most leaf page (P1) of a branch of the 
index tree (where a branch is an intermediate node in the index 
and all of the leaf nodes underneath it) with the left-most leaf 
page 
(P2) of the next two-level branch of the index tree, deleting 
leaf 
page P1 in the process 
  * is rolled back either explicitly via the ROLLBACK statement 
or because an error was hit 
- a SELECT statement is executed such that the SELECT: 
  * uses the UR (Unrepeatable Read) isolation level 
  * uses an index scan 
  * uses a start key that belongs in leaf page P1 (see above) 
before the page merge/deletion occurs 
  * starts probing the index while the page merge/deletion is in 
progress 
 
2.  To hit the problem in scenario #2 all of the following must 
be true: 
- a REORG INDEXES command is executed with the CLEANUP ONLY ALL 
option OR a REORG TABLE command is executed with the INPLACE 
option such that the REORG INDEXES/TABLE command: 
  * merges the right-most leaf page (P1) of a branch of the 
index tree with the left-most leaf page (P2) of the next 
branch of the index tree, deleting leaf page P1 
in the process 
  * hits a log-full error when writing the last log record that 
is part of the page merge/deletion operation 
- a SELECT statement is executed such that the SELECT: 
  * uses an index scan 
  * uses a start key that belongs in leaf page P1 (see above) 
before the page merge/deletion occurs 
  * starts probing the index while the page merge/deletion is in 
progress

**************************************************************** 
* USERS AFFECTED:                                              * 
* All users                                                    * 
**************************************************************** 
* PROBLEM DESCRIPTION:                                         * 
* See error description.                                       * 
**************************************************************** 
* RECOMMENDATION:                                              * 
* Upgrade to DB2 9.7.0.6.                                      * 
****************************************************************

DB2 Version 9.7 Fix Pack 6 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 7 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 8 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 9 for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 9a for Linux, UNIX, and Windows
 DB2 Version 9.7 Fix Pack 10 for Linux, UNIX, and Windows
 

The problem is first fixed in DB2 9.7.0.6.

not known / see Local fix

forerunner  : APAR is sysrouted TO one or more of the following: IC83426 
follow-up : 

Date  - problem reported    :
Date  - problem closed      :
Date  - last modified       :

24.02.2012
04.06.2012
04.06.2012

9.7.0.6