最近处理了一个问题,原因是由于命中ORA-600 [kole_t2u], [34] - description, bugs 导致 在udump 目录下大量转储 出cdmp 文件,
然后这些 cdmp 撑爆本地磁盘空间, 在oracle 发现本地无空间可写一些日志时,又触发ORA-7445 [_memmove()+64]
而触发ORA-600 [kole_t2u], [34] 的 根本原因是因为业务程序的非法操作,在db 中某张含有 clob 字段的表中插入了一些oracle现有字符集(zh16gbk)无法处理的字符
故障模拟如下:
SQL> select * from v$version;
BANNER
--------------------------------------------------------------------------------
Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production
PL/SQL Release 11.2.0.2.0 - Production
CORE 11.2.0.2.0 Production
TNS for Linux: Version 11.2.0.2.0 - Production
NLSRTL Version 11.2.0.2.0 - Production
SQL> create table t3(a clob);
Table created.
SQL> insert into t3 values(utl_raw.cast_to_varchar2('EC'));
insert into t3 values(utl_raw.cast_to_varchar2('EC'))
*
ERROR at line 1:
ORA-00600: internal error code, arguments: [kole_t2u], [34], [], [], [], [],[], [], [], [], [], []
SQL> select * from nls_database_parameters
PARAMETER VALUE
------------------------------ --------------------------------------------------------------------------------
NLS_LANGUAGE AMERICAN
NLS_TERRITORY AMERICA
NLS_CURRENCY $
NLS_ISO_CURRENCY AMERICA
NLS_NUMERIC_CHARACTERS .,
NLS_CHARACTERSET ZHS16GBK
NLS_CALENDAR GREGORIAN
NLS_DATE_FORMAT DD-MON-RR
NLS_DATE_LANGUAGE AMERICAN
NLS_SORT BINARY
NLS_TIME_FORMAT HH.MI.SSXFF AM
PARAMETER VALUE
------------------------------ --------------------------------------------------------------------------------
NLS_TIMESTAMP_FORMAT DD-MON-RR HH.MI.SSXFF AM
NLS_TIME_TZ_FORMAT HH.MI.SSXFF AM TZR
NLS_TIMESTAMP_TZ_FORMAT DD-MON-RR HH.MI.SSXFF AM TZR
NLS_DUAL_CURRENCY $
NLS_COMP BINARY
NLS_LENGTH_SEMANTICS BYTE
NLS_NCHAR_CONV_EXCP FALSE
NLS_NCHAR_CHARACTERSET AL16UTF16
NLS_RDBMS_VERSION 11.2.0.2.0
20 rows selected.
ORA-600 [kole_t2u], [34] 影响的版本范围:
Oracle Database - Enterprise Edition - Version 9.2.0.1 to 11.2.0.4 [Release 9.2 to 11.2]
Information in this document applies to any platform.
Oracle Server Enterprise Edition - Version: 9.2.0.1 to 11.1.0.6
***Checked for relevance on 20-Jan-2014***
ORA-600 [kole_t2u], [34] 触发的情景:
As stated, this error can come up in multibyte environments. Secondly it is clear that whenever this error comes up, there must be at least 1 incomplete codepoint used in the data.
In general we can split these occurrences in 3 categories:
Invalid multibyte data is being inserted by an application into a CLOB
Invalid multibyte data has been inserted in a VARCHAR2 (without initially being detected), and the stored data is moved to a CLOB at a later stage (either through application code, or by a Oracle process like Auditing).
Existing correctly stored CLOB data is incorrectly "split" into chunks. This could leave a codepoint "split" in the middle of the byte stream, leaving a incorrect number of bytes for the last codepoint before the split. This could either happen in application code, or could be due to bug in the database.
We will look at all these 3 categories in depth in the following paragraphs.
odm fund:
ORA-600 [kole_t2u], [34] - description, bugs, and reasons (文档 ID 734474.1)
ORA-600 [kole_t2u] With Multibyte Character Sets While Appending To LOB In a Loop (文档 ID 739282.1)
AL32UTF8 / UTF8 (Unicode) Database Character Set Implications (文档 ID 788156.1)
Character set conversion when using UTL_FILE (文档 ID 227531.1)
Getting ORA-7445 [_memmove()+64] and Instance Crashed (文档 ID 1294148.1) 转到底部转到底部
In this Document
Symptoms
Cause
Solution
This document is being delivered to you via Oracle Support's Rapid Visibility (RaV) process and therefore has not been subject to an independent technical review.
APPLIES TO:
Oracle Database - Enterprise Edition - Version 10.2.0.4 to 11.2.0.2 [Release 10.2 to 11.2]
Information in this document applies to any platform.
SYMPTOMS
In the alert log file there are reported errors like:
ORA-07445: exception encountered: core dump [_memmove()+64] [SIGBUS]
[Invalid address alignment] [0x0.........] [] []
and instance crashed.
The trace files generated by these errors are truncated (not relevant) or have no information (0 bytes).
CAUSE
In this case the issue is caused by a resource problem outside Oracle.
By checking the OS system log, you can see that the Oracle mount point has ran out of space just before the errors and until the instance crash.
For example in an HP-UX system, the messages can be as the following:
vmunix: vxfs: NOTICE: msgcnt 1 mesg 001: V-2-1: vx_nospace - /dev/vg01/lvora file system full (1 block extent)
SOLUTION
Monitor your system in order to avoid running out of space in Oracle mount points.
--转自
该贴由koei123转至本版2015-2-6 4:47:18