今天在处理Windows 平台的Oracle 10.2.0.3版本数据库升级到11.2.0.3版本时，系统出现以下的症状：

-- AlertLOG警告日志

Wed Dec 04 13:06:24 2013

Exception [type: ACCESS_VIOLATION, UNABLE_TO_READ] [ADDR:0x18] [PC:0x2E7368C, hshget1()+198]

Errors in file D:\APP\ADMINISTRATOR\diag\rdbms\zh10g\zh10g\trace\zh10g_ora_2488.trc  (incident=25447):

ORA-07445: 鍑虹幇寮傚父閿欒?: 鏍稿績杞?偍 [hshget1()+198] [ACCESS_VIOLATION] [ADDR:0x18] [PC:0x2E7368C] [UNABLE_TO_READ] []

Incident details in: D:\APP\ADMINISTRATOR\diag\rdbms\zh10g\zh10g\incident\incdir_25447\zh10g_ora_2488_i25447.trc

Use ADRCI or Support Workbench to package the incident.

See Note 411.1 at My Oracle Support for error and packaging details.

Wed Dec 04 13:06:26 2013

Dumping diagnostic data in directory=[cdmp_20131204130626], requested by (instance=1, osid=2488), summary=[incident=25447].

Wed Dec 04 13:06:28 2013

Sweep [inc][25447]: completed

Sweep [inc2][25447]: completed

-- INCIDENT日志

Dump continued from file: D:\APP\ADMINISTRATOR\diag\rdbms\zh10g\zh10g\trace\zh10g_ora_2488.trc

ORA-07445: 出现异常错误: 核心转储 [hshget1()+198] [ACCESS_VIOLATION] [ADDR:0x18] [PC:0x2E7368C] [UNABLE_TO_READ] []

========= Dump for incident 25447 (ORA 7445 [hshget1()+198]) ========

----- Beginning of Customized Incident Dump(s) -----

Exception [type: ACCESS_VIOLATION, UNABLE_TO_READ] [ADDR:0x18] [PC:0x2E7368C, hshget1()+198]

Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - 64bit Production

Process Id: 0x00000538  Thread Id : 0x000009b8    Time : Wed Dec 04 13:06:24 

Excp. Code: 0xc0000005  Excp. Type: ACCESS_VIO    Flags: 0x00000000

------------------- Registers ----------------------------

ip=0000000002E7368C sp=000000001430EDF0 rp=0000000000000001

r1=000007FF37A498A8 r2=0000000000000002 r3=0000000000000000 

r4=0000000000000002 r5=000000001430EDF0 r6=0000000000000001 r7=000007FF3767CA58 

r8=0000000000000000 r9=000007FF37A496D8 r10=0000000000000002 r11=000000000AEB3D78 

r12=0000000000000002 r13=000000000CB36F08 r14=0000000014313370 r15=0000000000000002 

------------------- End of Registers ---------------------

*** 2013-12-04 13:06:24.918

dbkedDefDump(): Starting a non-incident diagnostic dump (flags=0x3, level=3, mask=0x0)

----- Current SQL Statement for this session (sql_id=fyvqzrmdj8g9s) -----

ALTER PACKAGE "SYS"."DBMS_REGISTRY_SYS" COMPILE BODY REUSE SETTINGS

-- 数据库症状

SYS@ zh10g SQL>select open_mode from v$database;

OPEN_MODE

----------------------------------------

READ WRITE

SYS@ zh10g SQL>SELECT dbms_registry_sys.time_stamp('utlrp_bgn') as timestamp from dual

  2  ;

SELECT dbms_registry_sys.time_stamp('utlrp_bgn') as timestamp from dual

*

第 1 行出现错误:

ORA-03113: 通信通道的文件结尾

进程 ID: 2488

会话 ID: 5 序列号: 5

ERROR:

ORA-03114: 未连接到 ORACLE

解决方案：

ORA-07445: exception encountered: core dump [mdcsgetwkt()+111] [ACCESS_VIOLATION] [ADDR:0x2A8C1928] [PC:0x14691B067] [UNABLE_TO_READ] (Doc ID 1574679.1)

Applies to:

Symptoms

On Windows Patch bundles 20-22, creating a Spatial index or running a Spatial query fails with the following errors:

ORA-07445: exception encountered: core dump [mdcsgetwkt()+111] [ACCESS_VIOLATION] [ADDR:0x2A8C1928] [PC:0x14691B067] [UNABLE_TO_READ]

"no more data to read from socket"

The call stack in the trace file has the following function calls:

----- Call Stack Trace -----
mdcsgetwkt mdgsetgeocs mdidxs spefcpfa spefmccallstd pextproc psdexsp rpiswu2 psdextp pefccal 

Changes

Installation of Windows Patch bundle 20, 21, or 22

Cause

This issue was identified in the following bug:

Bug 17082161 : ON PATCH 20, 21, SPATIAL QUERIES FAIL WITH ORA-7445 [MDCSGETWKT()+111]
 
The root cause for the core dumps and "no more data to read from socket" errors is the backport fix (BLR) for Bug 16339328: SPATIAL: Performance enhancement to validate_geometry.  
 

Solution

Windows Patch bundle 23 has a new, corrected version of bug fix 16339328. To resolve this issue, install Windows Patch 23 (17163633)

- link patch32bit to download patch 17163633 for Microsoft Windows (32-Bit)

- link patch64bit to download patch 17163634 for Microsoft Windows (x64)

References