The problem is with detach partition processing.
To hit this problem all of the following conditions must be
true:
- Multi-node DPF configuration
- Database is shut down (deactivated), either explicitly or
implicitly (not explicitly activated and last connection closed)
on non-catalog node, and catalog node remains active. The
shutdown must occur during the first phase of the task's 2-phase
commit processing.
- At the same time as the database deactivation, the second,
asynchronous phase of ALTER TABLE DETACH PARTITION is running
for a table having a partition on the deactivating node. This
phase, called "APD", may occur somewhat after the DDL command
itself was executed, as it waits for its start conditions and
also to be scheduled within the ABP infrastructure. The start
conditions for the task include: the DETACH transaction has
committed, and any dependent MQT's required refresh has been
performed
As a result, the meta data for the detach partition may be
corrupted. At one or more database partitions, the detach
partition may think that it is still part of a partitioned
table.
This could result in several symptoms depending on what is then
executed on this table. Some example follow.
Example #1:
2012-03-21-14.57.26.346412-240 I7931E1690 LEVEL: Severe
PID : 3312 TID : 46912925329728PROC :
db2sysc 1
INSTANCE: db2inst1 NODE : 001 DB : SAMPLE
APPHDL : 1-77 APPID: *N1.db2inst1.120321185722
AUTHID : DB2INST1
EDUID : 51 EDUNAME: db2agent (SAMPLE) 1
FUNCTION: DB2 UDB, data management, sqldGetTableData, probe:1721
MESSAGE : ZRC=0x82040001=-2113667071=SQLD_NONSEVERE_PRGERR
"non-severe dms programming error"
DIA8532C An internal processing error has occurred.
DATA #1 : String, 71 bytes
Get table data called for partition table but global TCB wasn't
loaded.
Example #2:
2012-03-21-15.12.50.868836-240 I4260E1014 LEVEL: Error
PID : 6050 TID : 46912870803776PROC :
db2sysc 1
INSTANCE: db2inst1 NODE : 001 DB : SAMPLE
APPHDL : 0-77 APPID: *N0.db2inst1.120321191010
AUTHID : DB2INST1
EDUID : 117 EDUNAME: db2agntp (SAMPLE) 1
FUNCTION: DB2 UDB, lock manager, sqlpUpgradeLock, probe:520
MESSAGE : Cannot find my LRB! hash_p 0x2aaad402d150 tranptr
(nil) transChain
0x2aaad3b2ce90
DATA #1 : Hexdump, 96 bytes
0x00002AAAC0FFA828 : FAFF 0080 0000 0000 0000 0000 5400 0000
............T...
0x00002AAAC0FFA838 : 0000 0000 0000 0000 0000 0000 0000 0000
................
0x00002AAAC0FFA848 : 0000 0000 0000 0000 0000 0000 0000 0000
................
0x00002AAAC0FFA858 : 0000 0000 0000 0000 0100 0000 0000 0000
................
0x00002AAAC0FFA868 : 0000 0000 0000 0000 0000 0800 0000 0000
................
0x00002AAAC0FFA878 : FEFF FFFF 0000 0040 0000 0000 AA2A 0000
.......@.....*..
2012-03-21-15.12.50.882022-240 I5275E486 LEVEL: Severe
PID : 6050 TID : 46912870803776PROC :
db2sysc 1
INSTANCE: db2inst1 NODE : 001 DB : SAMPLE
APPHDL : 0-77 APPID: *N0.db2inst1.120321191010
AUTHID : DB2INST1
EDUID : 117 EDUNAME: db2agntp (SAMPLE) 1
FUNCTION: DB2 UDB, relation data serv, sqlrr_signal_handler,
probe:100
MESSAGE : DIA0505I Execution of a component signal handling
function has begun.
...
2012-03-21-15.12.50.885296-240 I11206E535 LEVEL: Severe
PID : 6050 TID : 46912870803776PROC :
db2sysc 1
INSTANCE: db2inst1 NODE : 001 DB : SAMPLE
APPHDL : 0-77 APPID: *N0.db2inst1.120321191010
AUTHID : DB2INST1
EDUID : 117 EDUNAME: db2agntp (SAMPLE) 1
FUNCTION: DB2 UDB, data management, sqldCreateIndex, probe:30
RETCODE : ZRC=0x8710001D=-2028994531=SQLP_LERR "Fatal Logic
Error"
DIA8526C A fatal error occurred in data protection
services.
Example #3:
2012-03-21-15.01.52.393301-240 I3354E1520 LEVEL: Severe
PID : 6050 TID : 46912925329728PROC :
db2sysc 1
INSTANCE: db2inst1 NODE : 001 DB : SAMPLE
APPHDL : 0-51 APPID: *N0.db2inst1.120321190055
AUTHID : DB2INST1
EDUID : 50 EDUNAME: db2agntp (SAMPLE) 1
FUNCTION: DB2 UDB, data management, sqldCheckTCBExistence,
probe:898
MESSAGE : ZRC=0x82040001=-2113667071=SQLD_NONSEVERE_PRGERR
"non-severe dms programming error"
DIA8532C An internal processing error has occurred.
DATA #1 : String, 36 bytes
Global TCB already exists on created |
