ORA-00600 [kcfnew_2], [202], [1702144], [12800], [12800]

os:aix7.1
db:oracle 11.2.0.3 rac

rdbms的alert日志中，从4月份起，m001进程不断的报错大量的600错误，该进程是mmon的slave进程，辅助进行性能诊断监控的，本文主要描述该故障的处理过程及方法。

1、具体的报错信息如下：
ORA-00600: internal error code, arguments: [kcfnew_2], [202], [1702144], [12800], [12800], [], [], [], [], [], [], []
2、查看trc文件，报错时执行的sql为：begin prvt_hdm.auto_execute( :dbid, :inst_num , :end_snap_id ); end;
3、查看trc文件中的call stack，发现有大量的temp方面的操作（kole_templob_init(),koklc_createtemp(),kdlt_createtemp()等），联系报错信息中有202，一般db_files默认为200，所以temp文件编号从200+fileid开始，202可能是代表fileid=2的临时表空间文件，所以猜想是temp表空间文件的问题。
4、查看dba_temp_files发现有临时表空间文件创建在本地。
5、通过600错误的函数在mos找到一个类似的错误EXPDP Failing With Ora-600 [kcfnew_2] (Doc ID 1509971.1)

This was caused by a corrupt segment in the temporary tablespace.
 
The following errors were seen in the alert.log shortly before the ORA-600 error:

alter database tempfile '/home/oradata/ora_dev/OBA08T/tempfile01' resize 1g
ORA-1516 signalled during: alter database tempfile '/home/oradata/ora_dev/OBA08T/tempfile01' resize 1g...

ORA-00600: internal error code, arguments: [kcfnew_2], [202], [947072], [619519], [619519],

Arg [a] FILE#
Arg [b] block to read
Arg [c] logical size kcflsz
Arg [d] physical size kcffsz

 

File# 202 belongs to the temporary table