Db2 engine abends with signal #11 (SIGSEGV) during a DROP TABLE
after a severe error message from sqlbPrepare2FreeExtOnTempSMP
in db2diag.log. The db2diag.log entries look similar to the
following:
2020-06-22-18.03.45.462189+120 I930184A659 LEVEL:
Severe
PID : 8192370 TID : 41545 PROC :
db2sysc
INSTANCE: XXXXXX NODE : 000 DB : XXXXX
APPHDL : 0-47874 APPID: XXXXXX
UOWID : 18 ACTID: 3
AUTHID : XXXXXX HOSTNAME: XXXXXX
EDUID : 41545 EDUNAME: db2agent (XXXXX)
FUNCTION: DB2 UDB, buffer pool services,
sqlbPrepare2FreeExtOnTempSMP, probe:173
DATA #1 :
Array boundary exceeded.
totalSmpPages: 1, smpPage: 1
sqlbGetLastInitSmpExtent(pdef): 0
pageSize: 32768, extentSize: 8, poolPage: 651536
2020-06-22-18.03.45.462698+120 I930844A667 LEVEL:
Severe
PID : 8192370 TID : 41545 PROC :
db2sysc
INSTANCE: XXXXXX NODE : 000 DB : XXXXX
APPHDL : 0-47874 APPID: XXXXXX
UOWID : 18 ACTID: 3
AUTHID : XXXXXX HOSTNAME: XXXXXX
EDUID : 41545 EDUNAME: db2agent (XXXXX)
FUNCTION: DB2 UDB, buffer pool services,
sqlbPrepare2FreeExtOnTempSMP, probe:196
DATA #1 : String, 37 bytes
smpMapInfo, no fixcb -- extent anchor
DATA #2 : Hexdump, 12 bytes
0x0A0000003AFEEE90 : 0000 0001 0000 07DA 0010 0000
............
The stack for the EDU in the FODC_Trap directory will look
similar to the following:
sqlbFreeExtentsForObject
sqlbDMSDeleteObject
sqldDropObj
sqldDropTable
sqldTableDrop
sqlrlDataPartDrop
sqlrldrt
sqlnq_drop_db_object_end
sqlnq_drop_db_object
This issue can occur when the follow happens:
1) There's a column organized table in a tablespace
2) An application has accessed the table
3) An extent movement (i.e. ALTER TABLESPACE ... REDUCE MAX) has
run and moved extents allocated to the table
4) An application does an uncommitted DROP TABLE on the table
and issues a ROLLBACK
5) An application does a DROP TABLE after the ROLLBACK has
completed |
