您的位置：首页 > 其它

ORA-01102: cannot mount database in EXCLUSIVE mode

2017-01-04 18:24 603 查看

ORA-01102:cannotmountdatabaseinEXCLUSIVEmode

安装完ORACEL10g数据库后,启动数据库时遇到ORA-01102:cannotmountdatabaseinEXCLUSIVEmode

[code]
[code][oracle@DB-Server~]$sqlplus/assysdba

SQL*Plus:Release10.2.0.1.0-ProductiononThuApr1013:53:082014

Copyright(c)1982,2005,Oracle.Allrightsreserved.

Connectedtoanidleinstance.

SQL>startup

ORACLEinstancestarted.

TotalSystemGlobalArea5033164800bytes

FixedSize2027672bytes

VariableSize905973608bytes

DatabaseBuffers4110417920bytes

RedoBuffers14745600bytes

ORA-01102:cannotmountdatabaseinEXCLUSIVEmode

[/code]
[/code]

查看告警日志文件，发现有ORA-09968:unabletolockfile和Linux-x86_64Error:11:Resourcetemporarilyunavailable等错误


[code]
[code][oracle@DB-Serverbdump]$tail-20falert_epps.log

SMONstartedwithpid=8,OSid=24482

RECOstartedwithpid=9,OSid=24484

CJQ0startedwithpid=10,OSid=24486

MMONstartedwithpid=11,OSid=24488

ThuApr1013:53:422014

startingup1dispatcher(s)fornetworkaddress'(ADDRESS=(PARTIAL=YES)(PROTOCOL=TCP))'...

MMNLstartedwithpid=12,OSid=24490

ThuApr1013:53:422014

startingup1sharedserver(s)...

ThuApr1013:53:422014

ALTERDATABASEMOUNT

ThuApr1013:53:422014

sculkget:failedtolock/u01/app/oracle//product/10.2.0/db_1/dbs/lkEPPSexclusive

sculkget:lockheldbyPID:20337

ThuApr1013:53:422014

ORA-09968:unabletolockfile

Linux-x86_64Error:11:Resourcetemporarilyunavailable

Additionalinformation:20337

ThuApr1013:53:422014

ORA-1102signalledduring:ALTERDATABASEMOUNT...

[/code]
[/code]

'

第一次遇到这种情况，于是网上搜索了一下资料，找到一篇相关文章http://blog.itpub.net/12272958/viewspace-716020，介绍了出现这种错误的三种情况：

a、Oracle的共享内存段或信号量没有被释放；

b、Oracle的后台进程（如SMON、PMON、DBWn等）没有被关闭；

c、用于锁内存的文件lk<sid>和sgadef<sid>.dbf文件没有被删除。

看了文章前面一部分。便武断的觉得应该是第三者情况（由于告警日志中出现了相关信息）：“用于锁内存的文件lk<sid>和sgadef<sid>.dbf文件没有被删除”。于是动手操作（太心急，这毛病老是难以改变），关闭数据库实例，删除/u01/app/oracle/product/10.2.0/db_1/dbs/lkEPPS文件后，重新启动数据库。结果作者的经历又在我身上重演了一次。

启动时候报ORA-00205错误


[code]
[code]SQL>shutdownimmediate;

ORA-01507:databasenotmounted

ORACLEinstanceshutdown.

SQL>startup

ORACLEinstancestarted.

TotalSystemGlobalArea5033164800bytes

FixedSize2027672bytes

VariableSize905973608bytes

DatabaseBuffers4110417920bytes

RedoBuffers14745600bytes

ORA-00205:errorinidentifyingcontrolfile,checkalertlogformoreinfo

[/code]
[/code]
查看告警日志，有如下错误信息：


[code]
[code]RECOstartedwithpid=9,OSid=24887

CJQ0startedwithpid=10,OSid=24889

MMONstartedwithpid=11,OSid=24891

ThuApr1014:04:052014

startingup1dispatcher(s)fornetworkaddress'(ADDRESS=(PARTIAL=YES)(PROTOCOL=TCP))'...

MMNLstartedwithpid=12,OSid=24894

ThuApr1014:04:062014

startingup1sharedserver(s)...

ThuApr1014:04:062014

ALTERDATABASEMOUNT

ThuApr1014:04:062014

ORA-00202:controlfile:'/u01/app/oracle/oradata/epps/control01.ctl'

ORA-27086:unabletolockfile-alreadyinuse

Linux-x86_64Error:11:Resourcetemporarilyunavailable

Additionalinformation:8

Additionalinformation:20341

ThuApr1014:04:092014

ORA-205signalledduring:ALTERDATABASEMOUNT...

[/code]
[/code]
于是索性静下心来多了解一下这个问题，在网上查看了大量关于ORA-01102的文章。这篇博客http://blog.itpub.net/20674423/viewspace-711545对这个问题阐述得比较清晰。下面是Metalink关于ORA-01102的解释以及解决方法：

ProblemDescription:

====================

Youaretryingtostartupthedatabaseandyoureceivethefollowingerror:

ORA-01102:cannotmountdatabaseinEXCLUSIVEmode

Cause:Someotherinstancehasthedatabasemountedexclusiveorshared.

Action:Shutdownotherinstanceormountinacompatiblemode.

ProblemExplanation:

====================

AdatabaseisstartedinEXCLUSIVEmodebydefault.Therefore,theORA-01102errorismisleadingandmayhaveoccurredduetooneofthefollowingreasons:

-thereisstillan"sgadef<sid>.dbf"fileinthe"ORACLE_HOME/dbs"directory

-theprocessesforOracle(pmon,smon,lgwranddbwr)stillexist

-sharedmemorysegmentsandsemaphoresstillexisteventhoughthe

databasehasbeenshutdown

-thereisa"ORACLE_HOME/dbs/lk<sid>"file

SearchWords:

=============

ORA-1102,crash,immediate,abort,fail,fails,migration

SolutionDescription:

=====================

Verifythatthedatabasewasshutdowncleanlybydoingthefollowing:

1.Verifythatthereisnota"sgadef<sid>.dbf"fileinthedirectory"ORACLE_HOME/dbs".

%ls$ORACLE_HOME/dbs/sgadef<sid>.dbfIfthisfiledoesexist,removeit.

%rm$ORACLE_HOME/dbs/sgadef<sid>.dbf

2.Verifythattherearenobackgroundprocessesownedby"oracle"

%ps-ef|grepora_|grep$ORACLE_SID

Ifbackgroundprocessesexist,removethembyusingtheUnix

command"kill".Forexample:

%kill-9<rocess_ID_Number>

3.Verifythatnosharedmemorysegmentsandsemaphoresthatareownedby"oracle"stillexist

%ipcs-b

Iftherearesharedmemorysegmentsandsemaphoresownedby"oracle",removethesharedmemorysegments

%ipcrm-m<Shared_Memory_ID_Number>

andremovethesemaphores

%ipcrm-s<Semaphore_ID_Number>

NOTE:Theexampleshownaboveassumesthatyouonlyhaveone

databaseonthismachine.Ifyouhavemorethanone

database,youwillneedtoshutdownallotherdatabases

beforeproceedingwithStep4.

4.Verifythatthe"$ORACLE_HOME/dbs/lk<sid>"filedoesnotexist

5.Startuptheinstance

SolutionExplanation:

=====================

The"lk<sid>"and"sgadef<sid>.dbf"filesareusedforlockingsharedmemory.Itseemsthateventhoughnomemoryisallocated,Oraclethinksmemoryisstilllocked.Byremovingthe"sgadef"and"lk"filesyouremoveanyknowledgeoraclehasofsharedmemorythatisinuse.Nowthedatabasecanstart.

参考资料：
http://blog.itpub.net/12272958/viewspace-716020http://blog.itpub.net/20674423/viewspace-711545

内容来自用户分享和网络整理，不保证内容的准确性，如有侵权内容，可联系管理员处理

标签：

相关文章推荐

新的分享

章节导航