DB2 FOR LUW / DB2FORLUW / 970 - DB2

If a crash recovery, log replay on HADR standby or rollforward 
operation is interrupted, the redo master (DB2 EDU) might result 
in corruption of a transaction log file. 
 
The following example shows a sample output produced when a 
crash recovery is interrupted : 
 
 
2011-07-07-14.18.22.754759-240 I287158306E539      LEVEL: Error 
PID     : 9501                 TID  : 47938059364672PROC : 
db2sysc 55 
INSTANCE:<instance-name>            NODE : 055          DB   : 
<db-name> 
APPHDL  : <apphdl>              APPID: <app id> 
AUTHID  : <authid> 
EDUID   : 246                  EDUNAME: db2redow (<dbname>) 55 
FUNCTION: DB2 UDB, recovery manager, sqlpPRecProcLog, probe:130 
RETCODE : ZRC=0x80100003=-2146435069=SQLP_INTRP "USER INTERRUPT 
DETECTED" 
          DIA8003C The interrupt  has been received. 
 
 
You can refer to the following messages in the db2diag.log file 
to confirm log file corruption: 
 
 
2011-07-07-14.18.22.772201-240 I287362638E418      LEVEL: 
Warning 
PID     : 9501                 TID  : 47938122279232PROC : 
db2sysc 55 
INSTANCE:<instance-name>            NODE : 055          DB   : 
<db-name> 
EDUID   : 1385                 EDUNAME: db2loggr (<dbname>) 55 
FUNCTION: DB2 UDB, recovery manager, sqlpgSwitchCopyPongToPing, 
probe:1040 
DATA #1 : <preformatted> 
writing BuffreArray 0 to log extent 1067 tailPage 17417 
 
2011-07-07-14.18.23.535697-240 I287505200E429      LEVEL: Info 
PID     : 9501                 TID  : 47938122279232PROC : 
db2sysc 55 
INSTANCE:<instance-name>            NODE : 055          DB   : 
<db-name> 
EDUID   : 1385                 EDUNAME: db2loggr (<dbname>) 55 
FUNCTION: DB2 UDB, recovery manager, sqlpgSwitchFromRedoToUndo, 
probe:824 
DATA #1 : <preformatted> 
TailIndex 0  firstLso 178309837281  nextLso 178380831814 
logGapSize 0 
 
2011-07-07-14.18.23.546895-240 I287505630E549      LEVEL: Severe 
PID     : 9501                 TID  : 47938118084928PROC : 
db2sysc 55 
INSTANCE:<instance-name>            NODE : 055 
EDUID   : 1389                 EDUNAME: db2loggw (<dbname>) 55 
FUNCTION: DB2 UDB, data protection services, sqlpgWriteToDisk, 
probe:909 
MESSAGE : ZRC=0x8610000D=-2045771763=SQLP_BADLOG "Log File 
cannot be used" 
          DIA8414C Logging can not continue due to an error. 
DATA #1 : <preformatted> 
TailPage 17417 does not match pagePso 178309837280 and firstLso 
178309837281 
 
2011-07-07-14.18.23.548124-240 I287507883E461      LEVEL: Error 
PID     : 9501                 TID  : 47938013227328PROC : 
db2sysc 55 
INSTANCE:<instance-name>            NODE : 055          DB   : 
<db-name> 
APPHDL  : <apphdl>              APPID: <app id> 
AUTHID  : <authid> 
EDUID   : 331                  EDUNAME: db2redom (<dbname>) 55 
FUNCTION: DB2 UDB, data protection services, sqlpgPostLoggw, 
probe:880 
MESSAGE : request 13 postcode -2045771763

**************************************************************** 
* USERS AFFECTED:                                              * 
* Users interrupting crash recovery                            * 
**************************************************************** 
* PROBLEM DESCRIPTION:                                         * 
* There is a rare timing hole in crash  recovery due to  which * 
* a transaction log file can be corrupted if the crash         * 
* recovery is interrupted during the very rare timing hole in  * 
* the logic which maintains the consistency of the log records * 
* in the transaction log file.                                 * 
*                                                              * 
* The issue could be avoided if no interrupt is issued during  * 
* crash recovery.  But, once experienced the database has to   * 
* be restored and rollforward could be done before the log     * 
* corruption.                                                  * 
**************************************************************** 
* RECOMMENDATION:                                              * 
* Users can upgrade to DB2 V9.7 fix pak 5 to avoid this        * 
* defect.                                                      * 
* Or, avoid interrupting crash recovery.                       * 
****************************************************************

Avoid interrupts such as forced system reboots during crash 
recovery. 
Once a transaction log is corrupted, the database must be 
restored from a backup and a roll forward operation to a PIT 
must be performed.

DB2 Version 9.7 Fix Pack 5 for Linux, UNIX, and Windows
 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
 

First Fixed in DB2 Version V9.7 fix pak 5

not known / see Local fix

forerunner  : APAR is sysrouted TO one or more of the following: IC77713 IC77714 
follow-up : 

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

13.07.2011
15.11.2011
15.11.2011

9.7.