基本信息，11.2.0.4的3节点rac，aix 6.1操作系统。

数据库2节点出现600错误后，数据库实例宕掉，其余2个节点正常，集群在问题期间也未发生重启，本次影响的仅仅是2节点实例的重启。

Fri Apr 05 17:28:38 2019

Errors in file /u01/app/oracle/diag/rdbms/db102/db1022/trace/db1022_lmon_3474368.trc  (incident=160089):

ORA-00600: internal error code, arguments: [kghstack_underflow_internal_2], [0x1108E6388], [], [], [], [], [], [], [], [], [], []

Incident details in: /u01/app/oracle/diag/rdbms/db102/db1022/incident/incdir_160089/db1022_lmon_3474368_i160089.trc

Fri Apr 05 17:28:42 2019

Dumping diagnostic data in directory=[cdmp_20190405172842], requested by (instance=2, osid=3474368 (LMON)), summary=[incident=160089].

Use ADRCI or Support Workbench to package the incident.

See Note 411.1 at My Oracle Support for error and packaging details.

Errors in file /u01/app/oracle/diag/rdbms/db102/db1022/trace/db1022_lmon_3474368.trc:

ORA-00600: internal error code, arguments: [kghstack_underflow_internal_2], [0x1108E6388], [], [], [], [], [], [], [], [], [], []

LMON (ospid: 3474368): terminating the instance due to error 481

System state dump requested by (instance=2, osid=3474368 (LMON)), summary=[abnormal instance termination].

System State dumped to trace file /u01/app/oracle/diag/rdbms/db102/db1022/trace/db1022_diag_5637018_20190405172846.trc

Instance terminated by LMON, pid = 3474368

分析时也寻找到了俩篇mos

LMON or LMS Process Crashes Instance With ORA-600 [kghstack_underflow_internal_2] (文档 ID 2003278.1)

Bug 18687067 : ORA-600 [KGHSTACK_UNDERFLOW_INTERNAL_2]

   老盖也发过一篇blog说是110204 aix平台下一个专有的bug，Bug的原因是 AIX C编译代码产生的BUG，他的那篇博客中的callstack跟上面提到的mos一致，但跟我遇到的情况也不一样。 结合问题节点采集的数据，callstack对应不上。

   请问是否能够分析出本次故障的原因。