ORA-00600: internal error code, arguments: [kcbz_check_objd_typ_3]

今天在做資料庫例檢時，alert.log檔案中出現如下錯誤：

Thu Jun 9 09:00:27 2011
Errors in file /u01/app/oracle/admin/xxxx/bdump/xxxx2_m001_577740.trc:
ORA-00600: internal error code, arguments: [kcbz_check_objd_typ_3], [0], [0], [1], [], [], [], []

查了一下，有如下BUG：

NB	Bug	Fixed	Description
	7251049	11.2.0.2, 12.1.0.0	Corruption in bitmap index introduced when using transportable tablespaces	V10020002 V10020003 V10020004 V10020005 V11010006 V11010007 V11020001 BITMAP CORR/IND ERROR OERI TRANSPORT WRONGRES
+	8597106	11.2.0.2, 12.1.0.0	Lost Write in ASM when normal redundancy is used	V11010006 V11010007 V11020001 ASM CORR CORR/PHY ERROR OERI
	9247416	11.2.0.2, 12.1.0.0	Corrupt bitmap index after transport using conventional exp / imp	V11020001 BITMAP CORR/DIC CORR/IND EXP TRANSPORT
+	10209232	11.1.0.7.7, 11.2.0.2.1, 12.1.0.0	ORA-1578 / ORA-600 [3020] Corruption. Misplaced Blocks and Lost Write in ASM	V11010006 V11010007 V11020001 V11020002 ASM CORR CORR/LOG CORR/PHY ERROR OERI
	6405339	10.2.0.5, 11.1.0.7, 11.2.0.1	Corruption / OERI[kcbz_check_objd_typ_3] / OERI:5400 / OERI:5357 from PQ	V10020002 V10020003 V10020004 V11010006 CORR OERI PQO
+	4430244	10.2.0.4, 11.1.0.6	Segment advisor can load blocks of dropped objects into buffer cache (KCB OERI errors)	V10020002 V10020003 DBMSPKG OERI
	4592596	10.2.0.4, 11.1.0.6	Corruption (ORA-1410 / ORA-8103) from multi-table insert with direct load	V09020008 V10010005 V10020002 V10020003 CORR/PHY DPATH MTI
	5689412	10.2.0.5	ORA-600 [kcbz_check_objd_typ_3] after SCN based recovery

/u01/app/oracle/admin/xxxx/bdump/xxxx2_m001_577740.trc
Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - 64bit Production
With the Partitioning, Real Application Clusters, OLAP and Data Mining options
ORACLE_HOME = /u01/app/oracle/10.2.0
System name: AIX
Node name: host2
Release: 1
Version: 6
Machine: 000xxxxxxx00
Instance name: xxxx2
Redo thread mounted by this instance: 2
Oracle process number: 105
Unix process pid: 577740, image: (m001)

*** ACTION NAME:(Auto ADDM Slave Action) 2011-06-09 09:00:18.243
*** MODULE NAME:(MMON_SLAVE) 2011-06-09 09:00:18.243
*** SERVICE NAME:(SYS$BACKGROUND) 2011-06-09 09:00:18.243
*** SESSION ID:(983.27208) 2011-06-09 09:00:18.242
*** SESSION ID:(983.27208) 2011-06-09 09:00:18.242
OBJD MISMATCH typ=6, seg.obj=67378, diskobj=66461, dsflg=100000, dsobj=67378, tid=67378, cls=1
Formatted dump of block:
buffer tsn: 2 rdba: 0x00c0d150 (3/53584)
scn: 0x0000.0451d963 seq: 0x01 flg: 0x02 tail: 0xd9630601
frmt: 0x02 chkval: 0x0000 type: 0x06=trans data
Hex dump of block: st=0, typ_found=1
Dump of memory from 0x0700000223902000 to 0x0700000223904000
700000223902000 06A20000 00C0D150 0451D963 00000102 [.......P.Q.c....]
700000223902010 00000000 01000000 0001039D 0451D95E [.............Q.^]

看起來是做ADDM時出現的錯誤，那麼，是哪個Segment呢？

SQL> select object_id, data_object_id, owner, object_name, object_type from dba_objects
where data_object_id in (66461, 67378) or object_id in (66461, 67378);
2
OBJECT_ID DATA_OBJECT_ID OWNER
---------- -------------- ------------------------------
OBJECT_NAME
--------------------------------------------------------------------------------
OBJECT_TYPE
-------------------
67378 67378 SYS
WRH$_ACTIVE_SESSION_HISTORY
TABLE PARTITION

SQL> select /*+ full(t) */ count(*) from WRH$_ACTIVE_SESSION_HISTORY t;
^Cselect /*+ full(t) */ count(*) from WRH$_ACTIVE_SESSION_HISTORY t
*
ERROR at line 1:
ORA-00600: internal error code, arguments: [kcbz_check_objd_typ_3], [2], [0],
[6], [], [], [], []

問題重現

先解決掉問題

SQL> alter system flush buffer_cache;

System altered.

SQL> select /*+ full(t) */ count(*) from WRH$_ACTIVE_SESSION_HISTORY t;

COUNT(*)
----------
106493

注意的是，RAC需要對每個節點都進行alter system flush buffer_cache;

問題解決了，但導致這個問題的原因還是未知，估計也是個BUG。這個Oracle的版本是10.2.0.1

[@more@]