In a Db2 pureScale environment, severe db2diag.log messages with
error tag SQLO_ADDR_CONFLICT may be infrequently observed. The
following is an example of the db2diag.log error messages which
could be seen:
2018-03-25-00.40.20.697505-300 I6463E1602 LEVEL:
Severe (OS)
PID : 13604 TID : 47404375130112 PROC :
db2rocm 1 [db2inst1]
INSTANCE: db2inst1 NODE : 001
HOSTNAME: host01
FUNCTION: DB2 UDB, SQO Memory Management, sqlocshr, probe:210
MESSAGE : ZRC=0x850F00C4=-2062614332=SQLO_ADDR_CONFLICT
"An address space conflict was detected attaching to
shared memory."
CALLED : OS, -, shmat OSERR: EINVAL
(22)
DATA #1 : Memory set handle, PD_TYPE_OSS_MEM_SET_HDL, 48 bytes
0x00007FFF5889BEB8 : 0000 0000 0200 0000 0000 0000 0200 0000
................
0x00007FFF5889BEC8 : 0000 0000 0000 0000 0000 0000 0000 0000
................
0x00007FFF5889BED8 : 0000 1B16 0000 0000 1D00 4419 0000 0100
..........D.....
DATA #2 : Memory set ID, PD_TYPE_OSS_MEM_SET_ID, 4 bytes
0x000000020003AFD0 : 4680 DB20
F..
DATA #3 : Pointer, 8 bytes
0x00002b1d3c000000
CALLSTCK: (Static functions may not be resolved correctly, as
they are resolved to the nearest symbol)
[0] 0x00002B1D38EE68A1 sqlocshr + 0xE61
[1] 0x0000000000413B07
_Z23sqlhaInternalDbmsAttachPA1_13__jmp_buf_tagP18SQLHA_SHMEM_HAN
DLE + 0xE7
[2] 0x00000000004160B9
_Z21sqlhaMonitorPartitionsmP23SQLO_PGRP_FILE_CONTENTS + 0xE39
[3] 0x0000000000418D59
_Z27sqlhaMonitorPartitionAndLogsmP23SQLO_PGRP_FILE_CONTENTS26_sq
lhaExpectedMonitorState + 0x419
[4] 0x000000000041010E rocmDB2Monitor + 0x82E
[5] 0x00000000004094F7 main + 0x18F7
[6] 0x00002B1D40949B15 __libc_start_main + 0xF5
[7] 0x0000000000407B39 db2rocm 1 [db2inst1] + 0x7B39
2018-03-25-00.40.20.699000-300 E8066E685 LEVEL:
Error
PID : 13604 TID : 47404375130112 PROC :
db2rocm 1 [db2inst1]
INSTANCE: db2inst1 NODE : 001
HOSTNAME: host01
FUNCTION: DB2 UDB, high avail services, sqlhaInternalDbmsAttach,
probe:230
RETCODE : ZRC=0x850F00C4=-2062614332=SQLO_ADDR_CONFLICT
"An address space conflict was detected attaching to
shared memory."
DATA #1 : unsigned integer, 2 bytes
1
DATA #2 : Database Partition Number, PD_TYPE_NODE, 2 bytes
1
DATA #3 : Database Partition Number, PD_TYPE_NODE, 2 bytes
1
DATA #4 : unsigned integer, 2 bytes
0
DATA #5 : unsigned integer, 2 bytes
0
DATA #6 : Boolean, 1 bytes
true
2018-03-25-00.40.20.699410-300 I8752E703 LEVEL:
Severe
PID : 13604 TID : 47404375130112 PROC :
db2rocm 1 [db2inst1]
INSTANCE: db2inst1 NODE : 001
HOSTNAME: host01
FUNCTION: DB2 UDB, high avail services, db2rocm_main, probe:1780
MESSAGE : Process must re-exec due to address space conflicts.
DATA #1 : ROCM Action, PD_TYPE_ROCM_ACTION, 271632 bytes
action->version: 1
action->actor->actorType: DB2
action->actor->actorID: 1
action->actor->instName: db2inst1
action->actor->hostname: NOT_POPULATED
action->actor->options: NONE
action->command: MONITOR
DATA #2 : String, 46 bytes
/usr/local/opt/db2/db2inst1/sqllib/adm/db2rocm
DATA #3 : String, 7 bytes
db2rocm
Even though these db2diag.log messages are labelled as severe
and error levels, they can be safely ignored as it means that
the operation failed with an address conflict due to ASLR and
will be retried. |
