今天在RAC 测试环境发现RAC 的node2 数据库没有起来，经过alert log 发现有大量的
Ora-600 的错误。
ORA-00600: internal error code, arguments: [KGHALO4], [0x394C34060], [], [], [], [], [], []

经过Metalink 发现 有这么多状况。

Known Issues:
Bug# 5736850 See Note 5736850.8
SGA corruption / crash from PQO bloom filter
Fixed: 10.2.0.4, 11.1.0.6

Bug# 5631836 See Note 5631836.8
OERI[kghalo4] can occur
Fixed: 10.2.0.4, 11.1.0.6

Bug# 5345999 See Note 5345999.8
Dump / OERI accessing V$SPPARAMETER
Fixed: 10.2.0.4, 11.1.0.6

Bug# 4890281 See Note 4890281.8
OERI[KGHALO4] can occur under qesblGetFilter
Fixed: 10.2.0.3, 11.1.0.6

Bug# 4414666 See Note 4414666.8
OERI[KGHALO4] can occur on NUMA
Fixed: 10.2.0.2, 11.1.0.6

Bug# 3949307 See Note 3949307.8
SGA memory corruption / OERI:KGHALO4 with high concurrency
Fixed: 9.2.0.7

由于版本是10.2.0.1.0，Oracle 建议可能解决有3种方法。
1： 重启node2 的Server 和所有服务
2： 修改init.ora 参数 _enable_NUMA_optimization=FALSE
3 ：bug 4414666

TIP: Click help for a detailed explanation of this page. Bookmark Go to End Subject: Bug 4414666 - OERI[KGHALO4] can occur on NUMA Doc ID: 4414666.8 Type: PATCH Modified Date : 24-SEP-2008 Status: PUBLISHED Bug 4414666 OERI[KGHALO4] can occur on NUMAThis note gives a brief overview of bug 4414666. The content was last updated on: 09-MAY-2008 Click here for details of each of the sections below. Affects: Product (Component) Oracle Server (Rdbms) Range of versions believed to be affected Versions < 11 Versions confirmed as being affected 10.2.0.1 Platforms affected Generic (all / most platforms affected) Fixed: This issue is fixed in 10.2.0.2 (Server Patch Set) 11.1.0.6 (Base Release) Symptoms: Related To: Internal Error May Occur (ORA-600) ORA-600 [KGHALO4] NUMA Related _ENABLE_NUMA_OPTIMIZATION Description ORA-600 [KGHALO4] can on a NUMA machine with ksmasgn in the call stack. Workaround: Add the following parameter to init.ora: _enable_NUMA_optimization=FALSE Or run with a larger shared_pool_size parameter.

果然是这个Bug，决定使用_enable_NUMA_optimization=FALSE

经检测发现问题解决。