Has anybody seen this before? If so... any ideas on cause and solution.
IBM was not able to determine the cause.

We have a Data Warehouse BI database, UDB v8.2 running on Dell Itanium
64 and Redhat Linux Advanced server 3.

It appears that a rogue query can cause the following errors in the
db2diag.log. When this error occurs, the database will hang at shutdown.
If the error is not present, the database shuts down fine. (We use the
string "TCB" as our detection string, grepping through the log to
determine if a problem has occurred.

The diag log entries look like this:

2008-05-26-08.44.03.363771-360 I4101E431 LEVEL: Error
PID : 7502 TID : 2305843009324952624PROC : db2agntp
(PIWD02) 0
INSTANCE: db2inst1 NODE : 000 DB : PIWD02
APPHDL : 0-1090 APPID: GA070474.I90A.02EF86143902
FUNCTION: DB2 UDB, runtime interpreter, sqlrisrt, probe:50
DATA #1 : Hexdump, 4 bytes
0x60000FFFFFFF0E48 : 0100 0F8B ....

2008-05-26-08.44.03.363801-360 I4533E431 LEVEL: Error
PID : 7987 TID : 2305843009324952624PROC : db2agntp
(PIWD02) 0
INSTANCE: db2inst1 NODE : 000 DB : PIWD02
APPHDL : 0-1090 APPID: GA070474.I90A.02EF86143902
FUNCTION: DB2 UDB, runtime interpreter, sqlrisrt, probe:50
DATA #1 : Hexdump, 4 bytes
0x60000FFFFFFF0E48 : 0100 0F8B ....

2008-05-26-08.44.03.917539-360 I4965E352 LEVEL: Severe
PID : 11707 TID : 2305843009324952624PROC : db2agent
(PIWD02) 0
INSTANCE: db2inst1 NODE : 000 DB : PIWD02
APPHDL : 0-1090
FUNCTION: DB2 UDB, data management, sqldRemoveCachedTableEntry, probe:2012
MESSAGE : DMS App Term: insert TCB still fixed.

2008-05-26-08.44.03.937526-360 I5318E329 LEVEL: Severe
PID : 11707 TID : 2305843009324952624PROC : db2agent
(PIWD02) 0
INSTANCE: db2inst1 NODE : 000 DB : PIWD02
APPHDL : 0-1090
MESSAGE : SQLD_TCB:
DATA #1 : String, 51 bytes
Temp Table(3:9)=<1090><BIADMINS>.TEMP (00003,00009)




Thanks very much,
Mark Ohman

DB2 & UDB email list listserv db2-l LazyDBA home page