发表于: 2007.10.20 19:44
分类: ORACLE , Bug
出处: http://yangtingkun.itpub.net/post/468/406766
---------------------------------------------------------------
这两天和bug有缘，又是一个10.2.0.3的bug。
后台alert文件中显示如下的错误信息：
*** ACTION NAME:(Main session) 2007-10-18 17:57:14.984
*** MODULE NAME:(PL/SQL Developer) 2007-10-18 17:57:14.984
*** SERVICE NAME:(tradedb.us.oracle.com) 2007-10-18 17:57:14.984
*** SESSION ID:(323.25333) 2007-10-18 17:57:14.984
Exception signal: 11 (SIGSEGV), code: 1 (Address not mapped to object), addr: 0x0, PC: [0x10252f160, opidsa()+480]
ksedmp: internal or fatal error
ORA-07445: 出现异常错误: 核心转储 [opidsa()+480] [SIGSEGV] [Address not mapped to object] [0x000000000] [] []
Current SQL statement for this session:
select col.*, com.Comments
from sys.all_tab_columns col,
sys.all_col_comments com
where col.owner = :owner
and col.table_name = :table_name
and com.Owner (+) = :Owner
and com.Table_Name (+) = :table_name
and com.Column_Name (+) = col.Column_Name
order by col.column_id
进一步观察trace文件，发现：
SO: 728110c60, type: 4, owner: 728005770, flag: INIT/-/-/0x00
(session) sid: 323 trans: 0, creator: 728005770, flag: (c0000041) USR/- BSY/-/-/-/-/-
DID: 0002-0019-0001F0A3, short-term DID: 0002-0019-0001F0A4
txn branch: 0
oct: 3, prv: 0, sql: 4ec95bf98, psql: 3f3d27958, user: 64/ANHUI_OPERATOR
O/S info: user: Administrator, term: ZJ-178, ospid: 1748:596, machine: WORKGROUPZJ-178
program: PLSQLDev.exe
application name: PL/SQL Developer, hash value=1190136663
action name: Main session, hash value=1773317990
last wait for 'SQL*Net message from client' blocking sess=0x0 seq=67 wait_time=3786 seconds since wait started=0
driver id=54435000, #bytes=1, =0
Dumping Session Wait History
for 'SQL*Net message from client' count=1 wait_time=3786
driver id=54435000, #bytes=1, =0
for 'SQL*Net message to client' count=1 wait_time=2
driver id=54435000, #bytes=1, =0
for 'library cache lock' count=1 wait_time=421
handle address=5a9590300, lock address=596897b08, 100*mode+namespace=12d
for 'library cache lock' count=1 wait_time=352
handle address=5a95690c0, lock address=5963b4650, 100*mode+namespace=12d
for 'library cache lock' count=1 wait_time=410
handle address=4f4ff4018, lock address=4b2fdfc68, 100*mode+namespace=12d
for 'library cache lock' count=1 wait_time=321
handle address=4ec95b768, lock address=4b2fdb6b0, 100*mode+namespace=12d
for 'library cache lock' count=1 wait_time=440
handle address=5aae08790, lock address=59641cf20, 100*mode+namespace=12d
for 'library cache lock' count=1 wait_time=415
handle address=5827a2888, lock address=59685a9f0, 100*mode+namespace=12d
for 'SQL*Net message from client' count=1 wait_time=5398
driver id=54435000, #bytes=1, =0
for 'SQL*Net message to client' count=1 wait_time=1
driver id=54435000, #bytes=1, =0
temporary object counter: 0
从trace文件上面的信息看，是PL/SQL Developer工具在查询数据字典时出现的错误。
在Metalink上找到相关的bug：Bug 5648872 - Dump (opidsa) from DESCRIBE of a cursor。
Oracle的描述为：
This problem is believed to be introduced in 10.2.0.3It is possible for clients connected to a 10.2.0.3 database to intermittently dump under opidsa when describing a cursor. Workaround: Flushing the shared pool can clear the problem for a period of time.从目前看，Oracle还没有计划在哪个版本去fixed这个bug。关于这个bug的消息的描述可以参考METALINK：Doc ID: Note:5648872.8
从目前发现的例子，都是PL/SQL Developer版本5连接10g数据库造成的，将Developer版本升级后，就不会出现这个问题。
re: ORA-7445(opidsa)错误 
py3f.yydg.com.cn) 2007-12-12 08:23:51.389
