oracle 9.2.0.4上的一个bug

今天一到公司telnet上report database server就发现一大堆ora-07445错误。

当时头皮就发麻，ora-07445和有名的ora-00600错误都是一些oracle 内核的bug。

去metalink上一查果然是bug。

metalink相关的文章如下：

ORA-7445 [KNPCDP2_CLEARSTATS()+12] IN A SNAPSHOT REFRESH

---

*** 10/06/03 02:15 am ***

TAR:

----

2769926.1

.

PROBLEM:

--------

Snapshot refresh fails with ORA-7445 [knpcdp2_ClearStats()+12]while updating

system.def$_destination. This has occurred since upgrade to 9.2.0.4.

.

DIAGNOSTIC ANALYSIS:

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

gdb of knpcdp2_ClearStats for offset +12 shows:

.

0x4000000002b8bfc8 < knpcdp2_ClearStats>: ldo 0x70(%sp),%sp

0x4000000002b8bfcc < knpcdp2_ClearStats+4>: copy %r26,%r31

0x4000000002b8bfd0 < knpcdp2_ClearStats+8>: ldo 0x27(%r26),%r26

0x4000000002b8bfd4 < knpcdp2_ClearStats+12>: stdby,e %r0,1(%r26)

0x4000000002b8bfd8 < knpcdp2_ClearStats+16>: ldo -0x27(%r26),%r26

0x4000000002b8bfdc < knpcdp2_ClearStats+20>: stdby,b,m %r0,8(%r26)

0x4000000002b8bfe0 < knpcdp2_ClearStats+24>: copy %r25,%r23

0x4000000002b8bfe4 < knpcdp2_ClearStats+28>: std %r0,0(%r26)

.

This indicates that it is possibly the +Olibcalls issue in knpcd.c (this has

not yet been recompiled).

.

.

WORKAROUND:

-----------

None

.

RELATED BUGS:

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

3065127

.

REPRODUCIBILITY:

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

Occurs repeatedly in this particular database at the customer's site. Not

reproduced in-house.

.

TEST CASE:

----------

None

.

STACK TRACE:

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

knpcdp2_ClearStats

knpcdps_PutSummary

knpc_rsum_RecordSummary

knpc_puicd

pevm_icd_call_commo

pfrrun

peicnt

kkxexe

opiexe

opiodr

rpidrus

skgmstack

rpidru

rpiswu2

rpidrv

rpidexe

kkjex1e

kkjsexe

kkjrdp

opirip

opidrv

sou2o

main

.

SUPPORTING INFORMATION:

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

pasdb_j017_6617.trc

alert_pasdb.log

.

24 HOUR CONTACT INFORMATION FOR P1 BUGS:

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

.

DIAL-IN INFORMATION:

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

.

IMPACT DATE:

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

.

*** 10/06/03 02:28 am ***

*** 10/06/03 02:29 am *** (CHG: Sta->16)

*** 10/06/03 02:36 am *** (CHG: Asg->NEW OWNER)

*** 10/06/03 02:41 am ***

*** 10/06/03 05:08 am ***

*** 10/06/03 05:11 am ***

*** 10/06/03 05:32 am ***

*** 10/07/03 12:52 am *** (CHG: Sta->36)

*** 10/07/03 12:52 am ***

 

 

记起来，昨天做过emc的链路升级，prod和report数据库都down了一次，而且时间还不短，snapshot没有刷新过去。不知道升级到10.2.0.3会如何。

看来oracle的bug还不少啊。

 

在此仅记录一下，不准备打patch，要搞就升级到10g。

 

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/9340716/viewspace-421327/，如需转载，请注明出处，否则将追究法律责任。

转载于:http://blog.itpub.net/9340716/viewspace-421327/