During rollforward processing (e.g. database rollforward, HADR
Standby), the instance may panic while redoing a FMT_EMP log
record, whose purpose is to redo the initialization of an EMP
(extent map page), a page responsible for tracking which extents
are allocated by a specific object (more information in DB2
Information Center, keyword "extent map page").
The problem can typically happen if the EMP update is followed
by extent movement which moves the EMP to a lower position, and
finally a table space reduction is issued, thus truncating the
overall size of the table space. The EMP update should be
skipped during a rollforward since the update no longer applies.
The problem is that we are not skipping the update, therefore
attempting to re-initialize an invalid (old) EMP.
The DB2 diagnostic log file will show evidence of attempting to
latch a page whose page number is too high for the table space.
The following example shows an attempt to latch page 26088,
whereas the maximum page for the table space is 26079:
2012-10-12-06.11.25.089631-240 I58495E576 LEVEL:
Error
PID : 26266 TID : 140212466345728 PROC :
db2sysc 0
INSTANCE: db2inst1 NODE : 000 DB :
SAMPLE
APPHDL : 0-8 APPID: *LOCAL.DB2.121012101122
HOSTNAME: localhost
EDUID : 51 EDUNAME: db2redow (SAMPLE) 0
FUNCTION: DB2 UDB, buffer pool services, sqlbLatchPoolRange,
probe:20
MESSAGE : ZRC=0x84020009=-2080243703=SQLB_RC_PG_NUM_ERR
"Page number exceeds max -8100"
DIA8709E Segmented table page number was too high.
<...skipping...>
2012-10-12-06.11.25.109231-240 I62470E11195 LEVEL:
Severe
PID : 26266 TID : 140212466345728 PROC :
db2sysc 0
INSTANCE: db2inst1 NODE : 000 DB :
SAMPLE
APPHDL : 0-8 APPID: *LOCAL.DB2.121012101122
HOSTNAME: localhost
EDUID : 51 EDUNAME: db2redow (SAMPLE) 0
FUNCTION: DB2 UDB, buffer pool services, sqlbLatchPoolRange,
probe:30
DATA #1 : Page ID, PD_TYPE_SQLB_PAGE_ID, 4 bytes
26088
DATA #2 : Page ID, PD_TYPE_SQLB_PAGE_ID, 4 bytes
26088
<...skipping...>
2012-10-12-06.11.25.110040-240 I73666E747 LEVEL:
Severe
PID : 26266 TID : 140212466345728 PROC :
db2sysc 0
INSTANCE: db2inst1 NODE : 000 DB :
SAMPLE
APPHDL : 0-8 APPID: *LOCAL.DB2.121012101122
HOSTNAME: localhost
EDUID : 51 EDUNAME: db2redow (SAMPLE) 0
FUNCTION: DB2 UDB, buffer pool services, sqlbLatchPoolRange,
probe:80
DATA #1 : String, 301 bytes
Current Map: Index out of bounds
Map entry size: 56
Map entries: 1
Map size: 1
Set Range Off MaxExtent MaxPage StartStr EndStr Adj
Conts
---- ----- --- --------- ------- -------- ------ ---
-----
[ 0] [ 0] 0 3259 26079 0 3259 0 1
(0)
The call stack file generated during the event will show
routines similar to:
sqlb_panic
sqlbLatchPoolRange
sqlbCommonWriteSetup
sqlbWritePage
sqlbufix
sqlbRedo
sqldmrdo
sqlpRecDbRedo
sqlpPRecProcLog
sqlpParallelRecovery |
****************************************************************
* USERS AFFECTED: *
* See Error Description *
****************************************************************
* PROBLEM DESCRIPTION: *
* See Error Description *
****************************************************************
* RECOMMENDATION: *
* Upgrade to DB2 Version 9.7 for Linux, Unix, and Windows Fix *
* Pack 8 *
**************************************************************** |
