[新手请教] oracle服务总是自动关闭 ORA-07445错误

Question

alert_chgajdb.log中显示错误如下：
ORA-07445: exception encountered: core dump [ACCESS_VIOLATION] [unable_to_trans_pc] [PC:0x7C94BD02] [ADDR:0xFF016AFF] [UNABLE_TO_READ] []chgajdb_j000_5208.trc显示错误如下（只截取一段）
Dump file d:\oracle\product\10.2.0\admin\chgajdb\bdump\chgajdb_j000_5208.trc
Mon Aug 01 10:50:02 2011
ORACLE V10.2.0.1.0 - Production vsnsta=0
vsnsql=14 vsnxtr=3
Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Production
With the Partitioning, OLAP and Data Mining options
Windows NT Version V5.2 Service Pack 2
CPU                 : 4 - type 586, 2 Physical Cores
Process Affinity    : 0x00000000
Memory (Avail/Total): Ph:121M/2047M, Ph+PgF:1223M/3943M, VA:1273M/2047M
Instance name: chgajdbRedo thread mounted by this instance: 1Oracle process number: 15Windows thread id: 5208, image: ORACLE.EXE
*** 2011-08-01 10:50:02.355
ksedmp: internal or fatal error
ORA-07445: exception encountered: core dump [ACCESS_VIOLATION] [unable_to_trans_pc] [PC:0x7C95A0D0] [ADDR:0x52004D] [UNABLE_TO_WRITE] []
Current SQL information unavailable - no session.
----- Call Stack Trace -----
calling              call     entry                argument values in hex      
location             type     point                (? means dubious value)     
-------------------- -------- -------------------- ----------------------------
7C95A0D0                      00000000             
7C8223EE             CALL???  00000000             
71A837E6             CALL???  00000000             
71A82EA7             CALLrel  71A837C9             
71A83DAB             CALLrel  71A82E9F             
71A83D81             CALLrel  71A83D90             
71B66DBD             CALLreg  00000000             
71B66D9B             CALLrel  71B66DA9             
71B66D65             CALLrel  71B66D84             
71B66D03             CALLrel  71B66D1D             
71B678E9             CALLrel  71B66C91             
71B67843             CALLrel  71B6788B             
71B67BFD             CALLrel  71B677DA             
_sldmGetHostName+17  CALL???  00000000             
_sldmInit+65         CALLrel  _sldmGetHostName+0   868F5F4 100
_ldmInit+318         CALLrel  _sldmInit+0          868F5F0 8668764 868F5F0 0 23C
                                                   868EF8C 0 658
_keltnfy+148         CALLrel  _ldmInit+0           91EEA30 0 3389E9CC 69F60E8
                                                   91EEA20 0 1 91EEA28
_kscnfy+1334         CALLreg  00000000             F 3424C62C
_ksucrp+1195         CALLrel  _kscnfy+0            F 3424C62C
_kkjrdp+281          CALLrel  _ksucrp+0            1
_opirip+1163         CALLrel  _kkjrdp+0            
_opidrv+857          CALLrel  _opirip+0            32 4 91EFEC0
_sou2o+45            CALLrel  _opidrv+0            32 4 91EFEC0
_opimai_real+227     CALLrel  _sou2o+0             91EFEB4 32 4 91EFEC0
_opimai+92           CALLrel  _opimai_real+0       3 91EFEEC
_BackgroundThreadSt  CALLrel  _opimai+0            
art@4+422                                          
7C824826             CALLreg  00000000             

Answer 1

某个参数没有设置好吧。
比如说：哪个内存，共享池之类的，设得比较小。
Oracle process number: 15
Process数达到15时，就自己关闭了么？

Answer 2

我最近也遇到同样的问题，请问楼主你是怎么解决的。
我这边报的是：
Thu Dec 08 14:39:09 2011
Errors in file t:\oracle\bdump\xxxdb_j001_1024.trc:
ORA-07445: exception encountered: core dump [ACCESS_VIOLATION] [_kpnfch+224] [PC:0xF986B4] [ADDR:0x84] [UNABLE_TO_READ] []

Answer 3

oracle 11 里 碰到过，不过似乎和你这个不太一样 官方解决方案是 alter session set star_transformation_enabled = temp_disable;
or
alter session set "_optimizer_use_cbqt_star_transformation" = false;

Answer 4

Applies to:Oracle Server - Enterprise Edition - Version: 9.2.0.1 and later   [Release: 9.2 and later ]
Information in this document applies to any platform.
***Checked for relevance on 17-Nov-2011***
RDBMS 9.2 or greater on Windows platforms
SymptomsThe following errors can be seen in alert.log:
ORA-07445: exception encountered : core dump [ACCESS_VIOLATION] [kpudcr2c+89] [PC:0x100CD3D1] [ADDR:0x19EC0000] [UNABLE_TO_WRITE] []
ORA-07445: exception encountered : core dump [ACCESS_VIOLATION] [unable_to_trans_pc] [PC:0x5A1113A] [ADDR:0xFFFFFFFFFFFFFFFF] [UNABLE_TO_READ]ORA-27300: OS system dependent operation:WaitForSingleObject failed with status: 0
ORA-27301: OS failure message: The operation completed successfully.
ORA-27302: failure occurred at: sssxcpttcs5
ChangesNone
CauseThis errorORA-07445: exception encountered : core dump [ACCESS_VIOLATION] [unable_to_trans_pc] [PC:0x5A1113A] [ADDR:0xFFFFFFFFFFFFFFFF] [UNABLE_TO_READ]
is typical for WIN out of memory and is commonly seen when INIT.ORA parameters are set to that of the DBCA starter DB.
 This problem appears as likely on 64 bit platforms as on 32 bitSolution1. Please refer to Note 342443.1 and apply latest patch bundle.2. Have sufficient physical memory on Server so that you can allocate more SGA/PGA to the database. Increase SGA_MAX_SIZE  and SGA_TARGET so that you can accommodate SGA pools.
 Restart the instance.
Increase INIT.ORA memory parameters and make sure following pools are set to recommended value i.e. 200M.
SHARED_POOL_SIZEJAVA_POOL_SIZE
STREAMS_POOL_SIZE10g中增加SGA_TARGET 吧