当前位置: DBARoad > 故障案例 > 文章正文

发现/app目录空间使用率激增，经检查发现，主要是由于bdump目录下的cdmp_*这类目录导致：

[/app/oracle/admin/CMPR/bdump]$ ls -ld cdmp_*
drwxr-x— 2 oracle dba 122880 Oct 27 10:29 cdmp_20081027102510
drwxr-x— 2 oracle dba 122880 Oct 27 10:34 cdmp_20081027102852
drwxr-x— 2 oracle dba 122880 Oct 27 10:39 cdmp_20081027103415
drwxr-x— 2 oracle dba 122880 Oct 27 10:43 cdmp_20081027103802
drwxr-x— 2 oracle dba 122880 Oct 27 10:54 cdmp_20081027104342
drwxr-x— 2 oracle dba 122880 Oct 27 10:59 cdmp_20081027105407
drwxr-x— 2 oracle dba 122880 Oct 27 11:04 cdmp_20081027105959
drwxr-x— 2 oracle dba 122880 Oct 27 11:15 cdmp_20081027111130
drwxr-x— 2 oracle dba 122880 Oct 27 11:22 cdmp_20081027111744
drwxr-x— 2 oracle dba 122880 Oct 27 11:26 cdmp_20081027112221
drwxr-x— 2 oracle dba 122880 Oct 27 11:31 cdmp_20081027112609
drwxr-x— 2 oracle dba 122880 Oct 27 11:37 cdmp_20081027113330
drwxr-x— 2 oracle dba 122880 Oct 27 11:54 cdmp_20081027114955
drwxr-x— 2 oracle dba 122880 Oct 27 12:13 cdmp_20081027120840
drwxr-x— 2 oracle dba 122880 Oct 27 13:12 cdmp_20081027130808
drwxr-x— 2 oracle dba 122880 Oct 27 13:25 cdmp_20081027132045
drwxr-x— 2 oracle dba 122880 Oct 27 14:25 cdmp_20081027142118
drwxr-x— 2 oracle dba 122880 Oct 27 14:30 cdmp_20081027142515
drwxr-x— 2 oracle dba 122880 Oct 27 14:38 cdmp_20081027143401
drwxr-x— 2 oracle dba 122880 Oct 27 15:10 cdmp_20081027150613
drwxr-x— 2 oracle dba 122880 Oct 27 15:25 cdmp_20081027152103
drwxr-x— 2 oracle dba 122880 Oct 27 15:33 cdmp_20081027152846
drwxr-x— 2 oracle dba 122880 Oct 27 15:46 cdmp_20081027154210
drwxr-x— 2 oracle dba 122880 Oct 27 15:55 cdmp_20081027155112
drwxr-x— 2 oracle dba 122880 Oct 27 16:13 cdmp_20081027160847
drwxr-x— 2 oracle dba 122880 Oct 27 16:18 cdmp_20081027161328
drwxr-x— 2 oracle dba 122880 Oct 27 16:37 cdmp_20081027163228
drwxr-x— 2 oracle dba 122880 Oct 27 16:44 cdmp_20081027164010
drwxr-x— 2 oracle dba 131072 Oct 27 16:49 cdmp_20081027164442
drwxr-x— 2 oracle dba 131072 Oct 27 17:03 cdmp_20081027165921
drwxr-x— 2 oracle dba 122880 Oct 27 17:18 cdmp_20081027171322
drwxr-x— 2 oracle dba 131072 Oct 27 17:44 cdmp_20081027173940
drwxr-x— 2 oracle dba 131072 Oct 27 17:48 cdmp_20081027174419
drwxr-x— 2 oracle dba 131072 Oct 27 18:41 cdmp_20081027183650

[/app/oracle/admin/CMPR/bdump]$ du -sk cdmp_20081105110516
121128 cdmp_20081105110516
[/app/oracle/admin/CMPR/bdump]$ ls -ld cdmp_* | wc -l
269
[/app/oracle/admin/CMPR/bdump]$

每个目录有近120M的大小，短短几天，就占用了30G的大小了。
检查alter log发现，每个cdmp_*目录生成，都是伴随着ORA-600错误产生的

Sat Nov 1 08:40:52 2008
Errors in file /app/oracle/admin/CMPR/udump/cmpr1_ora_4222.trc:
ORA-00600: internal error code, arguments: [kclchkinteg_18], [], [], [], [], [], [], []
Sat Nov 1 08:46:05 2008
Trace dumping is performing id=[cdmp_20081101084535]

关于ORA-00600：kclchkinteg_18，曾开过一个SR：

ERROR
———————–
Ora-00600: internal error code, arguments: [kclchkinteg_18], [], [], [], [], [],

ACTION PLAN:
=============

We suspect that you are hitting the following Bug : 3574872 .

Trying to insert a row into the index , scanning the BMB to see which one of
them has free space and then getting the block in CR mode and then check with
the LE for integrity checks where we find that the block corresponding to this
LE has already been written out when we are trying to write it out. This seems
to be a timing problem at the moment and can be workarounded using
” _gc_integrity_checks=false ”

So please use the following workaround and check whether the following is working
or not.

This problem is already fixed in 10.1 .

解决该ORA-600错误后，目前暂无cdmp_*目当生成

— The End —