ITPub博客

首页 > Linux操作系统 > Linux操作系统 > ORA-00600: internal error code, arguments: [kxfpg1sg:INV],

ORA-00600: internal error code, arguments: [kxfpg1sg:INV],

原创 Linux操作系统 作者:rocky_gao 时间:2012-07-20 20:09:30 0 删除 编辑

kill session后回滚时间太长,客户要求在OS 级杀spid,kill -9 spid 后,oracle突然重启。

重启过程很快,10几秒就完成了,以至于我杀完OS 进程重新登陆数据库都没有发现。

Fri Jul 20 06:14:28 2012
Stopping background process CJQ0
Fri Jul 20 06:15:32 2012
Errors in file /u01/app/admin/xxx/bdump/xxx_smon_9832.trc:
ORA-00600: internal error code, arguments: [kxfpg1sg:INV], [0xFFFFFFFF7A108208], [], [], [], [], [], []
Fri Jul 20 06:15:33 2012
Fatal internal error happened while SMON was doing active transaction recovery.
Fri Jul 20 06:15:33 2012
Errors in file /u01/app/admin/xxx/bdump/xxx_smon_9832.trc:
ORA-00600: internal error code, arguments: [kxfpg1sg:INV], [0xFFFFFFFF7A108208], [], [], [], [], [], []
SMON: terminating instance due to error 474
Fri Jul 20 06:15:33 2012
Errors in file /u01/app/admin/xxx/bdump/xxx_rbal_9867.trc:
ORA-00474: SMON process terminated with error
Instance terminated by SMON, pid = 9832
Fri Jul 20 06:15:35 2012
Starting ORACLE instance (normal) 

metalink 上找到一个相关的bug ,

Bug 6954722 - OERI[15709] in SMON in RAC / instance crash [ID 6954722.8]

Affects:

Product (Component) Oracle Server (Rdbms)
Range of versions believed to be affected Versions BELOW 11.2
Versions confirmed as being affected
Platforms affected Generic (all / most platforms affected)

Fixed:

This issue is fixed in

Description

SMON may fail with ORA-600 [15709] crashing the instance.
This can happen in RAC as a result of a deadlock between
parallel transaction recovery slaves and the master SMON.
 
Workaround
set FAST_START_PARALLEL_ROLLBACK = false

另外一篇文章 ORA-600 [kxfpg1sg:INV] Leading To Instance Crash [ID 1467729.1],这样写

Cause

For the ora-600 kxfpg1sg:INV error there are a couple of bugs reporting similar behavior, but closed with no solution as issue was not reproduced.
The error is not critical, it causes database to crash as per SMON hitting an ORA-600, but introduces no corruption nor data loss.

Not identified the root cause, but verified the area where it comes from.
 
The SMON crash is reported while doing transaction recovery (in parallel in this case). So the smon error is not a cause but a consequence.
SMON received error while trying to allocate parallel server slave to run parallel recovery.
So setting "fast_start_parallel_rollback = false" will probably stop SMON from getting this ORA-600, since it won't use parallel slave, but this won't fix the root error: the ORA-600.

Solution


First option 1.Setting "fast_start_parallel_rollback = false" will stop SMON from getting this ORA-600, since it won't use parallel slave, but this won't fix the root error
OR
second option 2. Increase the number of parallel_servers and see if same behaviour is seen
 

关于fast_start_parallel_rollback参数:Fast_start_parallel_rollback: 指执行并行回退时确定进程的最大数量。
false:禁止并行回滚功能
low:2*cpu个slave进程数
high:4*cpu个slave进程数

来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/11976525/viewspace-736180/,如需转载,请注明出处,否则将追究法律责任。

请登录后发表评论 登录
全部评论

注册时间:2010-04-09

  • 博文量
    112
  • 访问量
    273439