Oracle-07445[kgghash]:Oracle BUG导致更新LOB字段时进程被KILL掉

原创 2014年01月07日 19:39:23
Oracle-07445[kgghash]:Oracle BUG导致更新LOB字段时进程被KILL掉

Mon Apr 22 14:49:58 2013
Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x2B1733615000] [PC:0x777EE03, kgghash()+367]
Errors in file /oracle/app/11gR1/diag/rdbms/portal/portal2/trace/portal2_ora_19188.trc (incident=186505):
ORA-07445: 出现异常错误: 核心转储 [kgghash()+367] [SIGSEGV] [ADDR:0x2B1733615000] [PC:0x777EE03] [Address not mapped to object] [] <<<<<<<<<
Incident details in: /oracle/app/11gR1/diag/rdbms/portal/portal2/incident/incdir_186505/portal2_ora_19188_i186505.trc
Mon Apr 22 14:55:08 2013
Thread 2 advanced to log sequence 4042
 Current log# 12 seq# 4042 mem# 0: +portalDG/portal/onlinelog/group_12.292.796513605
Mon Apr 22 14:55:08 2013
SUCCESS: diskgroup ARCHDG was mounted
Mon Apr 22 14:55:14 2013
SUCCESS: diskgroup ARCHDG was dismounted
Mon Apr 22 15:00:01 2013
Process 0x0x2147ed628 appears to be hung while dumping                <===进程被KILL掉
Current time = 1955749344, process death time = 1955689185 interval = 60000
Attempting to kill process 0x0x2147ed628 with OS pid = 19188
OSD kill succeeded for process 0x2147ed628
同时段的alert日志中确实有KILL process信息,并且在process被KILL前,该进程有ORA-07445错误报出,估计是Oracle的BUG。

Dump file /oracle/app/11gR1/diag/rdbms/portal/portal2/incident/incdir_186505/portal2_ora_19188_i186505.trc
Oracle Database 11g Enterprise Edition Release - 64bit Production
With the Partitioning, Real Application Clusters, OLAP, Data Mining
and Real Application Testing options
ORACLE_HOME = /oracle/app/11gR1/db
System name: Linux
Node name: PORTAL02
Release: 2.6.18-194.el5
Version: #1 SMP Tue Mar 16 21:52:39 EDT 2010
Machine: x86_64
Instance name: portal2
Redo thread mounted by this instance: 2
Oracle process number: 213
Unix process pid: 19188, image: oracle@PORTAL02

*** 2013-04-22 14:49:58.322
*** SESSION ID:(1403.45735) 2013-04-22 14:49:58.322
*** CLIENT ID:() 2013-04-22 14:49:58.322
*** SERVICE NAME:(portal) 2013-04-22 14:49:58.322
*** MODULE NAME:(JDBC Thin Client) 2013-04-22 14:49:58.322
*** ACTION NAME:() 2013-04-22 14:49:58.322
Dump continued from file: /oracle/app/11gR1/diag/rdbms/portal/portal2/trace/portal2_ora_19188.trc
ORA-07445: 出现异常错误: 核心转储 [kgghash()+367] [SIGSEGV] [ADDR:0x2B1733615000] [PC:0x777EE03] [Address not mapped to object] []

========= Dump for incident 186505 (ORA 7445 [kgghash()+367]) ========
----- Beginning of Customized Incident Dump(s) -----
Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x2B1733615000] [PC:0x777EE03, kgghash()+367]
%rax: 0x000000001605399e %rbx: 0x0000000000000000 %rcx: 0x000000009df422ed
%rdx: 0x00000000895e2309 %rdi: 0x00002b1733615000 %rsi: 0x00000000096d62a0
%rsp: 0x00007fffb42e8a20 %rbp: 0x00007fffb42e8a30 %r8: 0x0000000000000000
 %r9: 0x00000000c55f2e0a %r10: 0x000000000777ee03 %r11: 0x000000002eeb4000
%r12: 0x0000000219382310 %r13: 0x0000000125d120e0 %r14: 0x00002b1733541e50
%r15: 0x0000000160c68fd0 %rip: 0x000000000777ee03 %efl: 0x0000000000010293
 kgghash()+356 (0x777edf8) movzbl 0xa(%rdi),%esi
 kgghash()+360 (0x777edfc) shl $0x10,%esi
 kgghash()+363 (0x777edff) add %esi,%eax
 kgghash()+365 (0x777ee01) jmp 0x777edb7
> kgghash()+367 (0x777ee03) movzbl (%rdi),%esi
 kgghash()+370 (0x777ee06) add %esi,%edx
 kgghash()+372 (0x777ee08) jmp 0x777ed19
 kgghash()+377 (0x777ee0d) add (%rdi),%edx
 kgghash()+379 (0x777ee0f) add 0x4(%rdi),%ecx

*** 2013-04-22 14:49:58.339
----- Current SQL Statement for this session (sql_id=bb8bnr20h93mk) -----
update oa9.sys_process set fd_detail=:1 , fd_descriptor=:2 where fd_id=:3

----- Call Stack Trace -----
skdstdst <- ksedst1 <- ksedst <- dbkedDefDump <- ksedmp
 <- ssexhd <- restore_rt <- kgghash <- kkocsCreateBindSet <- kkocsInitAdaptiveCtx
 <- apadrv <- opitca <- PGOSF385_kksFullTypeCheck <- rpiswu2 <- kksSetBindType
 <- kksfbc <- opiexe <- kpoal8 <- opiodr <- ttcpip
 <- opitsk <- opiino <- opiodr <- opidrv <- sou2o
 <- opimai_real <- ssthrdmain <- main <- libc_start_main <- start

Cursor#35(0x2b1732bb2c28) state=BOUND curiob=0x2b1733541e50
 curflg=4c fl2=0 par=(nil) ses=0x214cd7880
----- Dump Cursor sql_id=bb8bnr20h93mk xsc=0x2b1733541e50 cur=0x2b1732bb2c28 -----
Dump Parent Cursor sql_id=bb8bnr20h93mk phd=0x21b32bee8 plk=0x21a12acb8
 sqltxt(0x21b32bee8)=update oa9.sys_process set fd_detail=:1 , fd_descriptor=:2 where fd_id=:3
 parent=0x1e7cda730 maxchild=01 plk=0x21a12acb8 ppn=n
cursor instantiation=0x2b1733541e50 used=1366613397 exec_id=0 exec=0
 child#0(0x218b3bb40) pcs=0x1a2dff738
 clk=0x21ac20ed8 ci=0x1e737a828 pn=0x21bd41568 ctx=0x125d120e0
 kgsccflg=0 llk[0x2b1733541e58,0x2b1733541e58] idx=0
 xscflg=130672 fl2=11100008 fl3=2222600 fl4=100
----- Bind Byte Code (IN) -----
 Opcode = 14 Bind Skip Long
Opcode = 2 Bind Twotask Scalar Sql In (may be out) Copy
oacdef = 0x1a3438220 Offsi = 48, Offsi = 32
 Opcode = 2 Bind Twotask Scalar Sql In (may be out) Copy
oacdef = 0x1a3438250 Offsi = 48, Offsi = 64
 Opcode = 1 Unoptimized
Null Offset
----- Bind Info (kkscoacd) -----
 oacdty=01 mxl=4001(5018) mxlc=00 mal=00 scl=00 pre=00
 oacflg=03 fl2=1000010 frm=01 csi=852 siz=4000 off=0
 kxsbbbfp=2b1733614060 bln=4000 avl=4001 flg=25
 value=" WZ3+"...
 oacdty=01 mxl=2000(1620) mxlc=00 mal=00 scl=00 pre=00
 oacflg=03 fl2=1000010 frm=01 csi=852 siz=2128 off=0
 kxsbbbfp=2b17336137d0 bln=2000 avl=822 flg=05
 value="<process parent="oa">
<initial-node id="N1" name="开始" nextTargetNodeIds="N2" parent="oaInitNode">
<result priority="0" target="N2"/>
<node id="N2" name="提交" nextTargetNodeIds="N4" parent="oaDraftNode">
 oacdty=01 mxl=128(64) mxlc=00 mal=00 scl=00 pre=00
 oacflg=03 fl2=1000010 frm=01 csi=852 siz=0 off=2000
 kxsbbbfp=2b1733613fa0 bln=128 avl=32 flg=01
 Frames pfr (nil) siz=0 efr (nil) siz=0
 Cursor frame dump
 enxt: 2.0x00000060 enxt: 1.0x00000898
 pnxt: 1.0x00000060
 kxscphp=0x2b1733606828 siz=1008 inu=176 nps=96
 kxscbhp=0x2b17335a5660 siz=11456 inu=10696 nps=10512

ORA-07445通常是ORACLE的BUG引起的,在文档ORA-7445 [kgghash] [ID 310172.1]中发现有符合我们现象的BUG:
Bug 13463131 Dump (kgghash) from bind peeking,BUG 13463131的完整内容:
Bind peeking may dump in kgghash() when long character datatypes are
being used.
The stack will typically include the following calls:
<===== 与我们的CASE相符合,表 oa9.sys_process有CLOB字段,且在我们的CASE中是call stack为kgghash <- kkocsCreateBindSet <- kkocsInitAdaptiveCtx

The bind variable that causes the dump shows the avl value is greater than
the bln (in the cursor dump in the trace).

   oacdty=96 mxl=4001(22518) mxlc=00 mal=00 scl=00 pre=00
   oacflg=01 fl2=1000010 frm=01 csi=873 siz=4000 off=0
   kxsbbbfp=0e42f060  bln=4000  avl=4001  flg=25
This situation can happen, for example, using a CLOB column and having
the client bind with CHAR.
<====抽取我们的dump文件相关内容,可以发现bln=4000 avl=4001,符合文档描述。
 oacdty=01 mxl=4001(5018) mxlc=00 mal=00 scl=00 pre=00
 oacflg=03 fl2=1000010 frm=01 csi=852 siz=4000 off=0
 kxsbbbfp=2b1733614060 bln=4000 avl=4001 flg=25 <<<<<<
 value=" WZ3+"...
  Disable bind peeking (set "_optim_peek_user_binds" = false)
  but note that this may cause changes to execution plans so
  may not be practical.
<=====Oracle给出的解决方案是禁用bind peeking,但是带来的风险是可能导致SQL执行计划改变
另外可以升级至11.2.0.3之后再apply patch 13463131解决。
Note: This fix supersedes fixes of bug 7352414 , bug 8763922.

What is "Bind Peeking" ?
With 9i, a new feature "Bind Peeking" was introduced to try to get around the issues associated with the selectivity of a bind variable.
During hard parsing of a query using bind variables, the binds are "peeked at" and selectivity is generated based on the bind and the underlying column statistics.
This method is successful but if the bind selectivity of the initial execution of a statement varied from the selectivity of subsequent executions with different sets of binds this could sometimes lead to plans being enerated that were not representative of the general query usage and which performed poorly.

This feather is controlled by the hidden parameter "_OPTIM_PEEK_USER_BINDS". the default value is 'true'.

Peeking is enabled if all of the following conditions are met:
 * - optimizer_features_enabled is greater than 8.2
 * - there are user binds
 * - the cursor is fully bound
 * - the cursor is not remote mapped
 * - the client is using V8OCI
 * - it's a select or DML
 * - peeking is possible for this statement and the statement is being
 * loaded for the first time, or the original cursor was peeked and
 * this cursor is being reloaded

For more details, please refer to:

Document 387394.1 Query using Bind Variables is suddenly slow
Document 430208.1 Bind Peeking By Example
"_OPTIM_PEEK_USER_BINDS" [Hidden] Reference Note [ID 201672.1]


Oracle BUG导致实例宕机:ORA-07445

现象: 客户的数据库(RAC环境:发生了实例异常宕机现象,伴随有ORA-07445错误: Sun Jun 23 01:00:06 2013 Exception [type: S...

Oracle LOB字段处理工具

  • 2015年06月05日 10:38
  • 42KB
  • 下载

ORA-07445: 出现异常错误: 核心转储(oracle bug)


通过JDBC操纵Oracle LOB字段

在Oracle中,LOB(Large Object,大型对象)类型的字段现在用得越来越多了。因为这种类型的字段,容量大(最多能容纳4GB的数据),且一个表中可以有多个这种类型的字段,很灵活,适用于数据...


有些时候产线DB进行保养,限制的时间比较短,而通常这个时候又有大量用户远程连接的情况下,通过shutdown immediate来关闭数据库,需要的时间比较长,虽然通过shutdown abort来关...
  • chisky
  • chisky
  • 2011年01月06日 13:47
  • 1431

java oracle setblob lob字段的问题

oracle jdbc setblob 增删改查


Oracle中 lob字段采用独立的lob segment来存储,因此表的大小不能只查看dba_segments. mos给出了查询语句: How to Compute the Siz...

Oracle DBLink 访问Lob 字段 ORA-22992 解决方法

这篇测试一下通过DBLink 访问含有Blob字段表的方法。 关于DBLINK 和 Lob 的说明,参考如下链接: Oracle DBLink

J2EE程序中使用oracle数据库LOB字段的总结(elathen)     最近在J2EE的项目中需要使用LOB字段保存文本信息以及图片和文件,到网上搜拉一下,还不少,仔细看拉一下,但...

Oracle Kill被锁进程

  • 2013年06月07日 13:49
  • 2KB
  • 下载
您举报文章:Oracle-07445[kgghash]:Oracle BUG导致更新LOB字段时进程被KILL掉