ora-00600: internal error code, arguments: [KGHLKREM1], [0x838000020], [], [], [], [], [], [], [], [(六)

--------------------- Binary Stack Dump ---------------------

再往前查看alertlog，发现还报了ora-07445

Tue Jan 17 08:42:12 2012
Archived Log entry 7472 added for thread 1 sequence 8444 ID 0x263e89b dest 1:
Tue Jan 17 09:00:14 2012
Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x8] [PC:0xB0997A, ksmdscan_internal()+82] [flags: 0x0, count: 1]
Errors in file /oracle/app/diag/rdbms/skate01/skate01/trace/skate01_ora_25574.trc (incident=264155):
ora-07445: exception encountered: core dump [ksmdscan_internal()+82] [SIGSEGV] [ADDR:0x8] [PC:0xB0997A] [Address not mapped to objec
t] []
Incident details in: /oracle/app/diag/rdbms/skate01/skate01/incident/incdir_264155/skate01_ora_25574_i264155.trc
Use ADRCI or Support Workbench to package the incident.
See Note 411.1 at My Oracle Support for error and packaging details.
Tue Jan 17 09:00:21 2012
Dumping diagnostic data in directory=[cdmp_20120117090021], requested by (instance=1, osid=25574), summary=[incident=264155].
Tue Jan 17 09:00:22 2012
Sweep [inc][264155]: completed
Sweep [inc2][264155]: completed
Tue Jan 17 09:06:08 2012
Media Recovery Waiting for thread 1 sequence 8446

然后查看oracle文档“ID 1070812.1”，发现这个我启用hugepage有关，

当系统vm.drop_caches设置大于0，并且启用hugepage，这时这两个就会冲突，因为drop_caches是要释放内存，而hugepage是hold住内存。

参考：http://blog.csdn.net/wyzxg/article/details/7279986

解决方法

1.如果启用hugepage，那就设置vm.drop_caches=0
[root@localhost ~]# more /proc/sys/vm/drop_caches
3
[root@localhost ~]# sysctl -a | grep drop_caches
vm.drop_caches = 3
[root@localhost ~]# vi /etc/sysctl.conf

##skate add
vm.drop_caches=0

使其立刻生效
[root@localhost ~]# sysctl -p

检查是否生效
[root@localhost ~]# sysctl -a | grep drop_caches
vm.drop_caches = 0

或者
2.升级Linux Kernel version到 2.6.18-194.0.0.0.4.EL5

附上官方文档：

ORA-600 [KGHLKREM1] On Linux Using Parameter drop_cache On hugepages Configuration [ID 1070812.1]

Oracle Server - Enterprise Edition - Version: 10.2.0.1 and later [Release: 10.2 and later ]
Generic Linux

You are running an Oracle Database, single-instance or RAC. You have the SGA backed by hugepages.

You are getting the error

and the SGA heap Dump of memory around the offending addr (in this particular example: 0x6bc00020)
it's showing zeroed out :

1. On your system you are running with vm.drop_caches=1 (or 3), drop_cache have been set to a value greater than zero , or you are executing

2. You have setup the Hugepages

This is a Linux Kernel issue.
Using the linux kernel "drop_cache" parameter and having the hugepages a memory corruption can occurs.

Per internal Bug 9461825, executing vm.drop_caches corrupts Oracle Database SGA hugepages;
it is fixed in Linux Kernel version 2.6.18-194.0.0.0.4.EL5

1. As a workaround when hugepages are set avoid any vm.drop_cache settings.

OR

2. Upgrade to Linux Kernel version 2.6.18-194.0.0.0.4.EL5

----------end-----------