O/S-Error: (OS 33) 另一个程序已…

itpub上看到个求助的帖子，错误比较典型，贴出来的alert log及后面有分析思路和参考文档

错误摘抄：
ALTER TABLESPACE "TOOLS" END BACKUP
Completed: ALTER TABLESPACE "TOOLS" END BACKUP
Mon Sep 01 19:28:11 2008
ALTER TABLESPACE "USERS" BEGIN BACKUP
Completed: ALTER TABLESPACE "USERS" BEGIN BACKUP
Mon Sep 01 19:30:56 2008
ALTER TABLESPACE "USERS" END BACKUP
Mon Sep 01 19:30:56 2008
Completed: ALTER TABLESPACE "USERS" END BACKUP
Mon Sep 01 19:30:56 2008
ALTER TABLESPACE "XDB" BEGIN BACKUP
Completed: ALTER TABLESPACE "XDB" BEGIN BACKUP
Mon Sep 01 19:30:57 2008
ALTER TABLESPACE "XDB" END BACKUP
Completed: ALTER TABLESPACE "XDB" END BACKUP
Mon Sep 01 19:30:57 2008

...(从上面的看，备份不是rman,而是把表空间置于backup状态，然后由备份软件copy，这个容易出问题，copy过程中就不能写文件)
ALTER DATABASE BACKUP CONTROLFILE TO 'D:\ORACLE\ORA92\DATABASE\BECFBARCODE.BAK' REUSE
Mon Sep 01 19:31:28 2008
Completed: ALTER DATABASE BACKUP CONTROLFILE TO 'D:\ORACLE\OR
Mon Sep 01 21:44:32 2008
Errors in file e:\barcode\admin\barcode\bdump\barcode_lgwr_3040.trc:
ORA-00345: redo log write error block 20528 count 2
ORA-00312: online log 3 thread 1: 'E:\BARCODE\ORADATA\BARCODE\REDO03.LOG'
ORA-27072: skgfdisp: I/O error
OSD-04008: WriteFile() 失败, 无法写入文件
O/S-Error: (OS 33) 另一个程序已锁定文件的一部分，进程无法访问。

。。。

Database Characterset is ZHS16GBK
replication_dependency_tracking turned off (no async multimaster replication found)
Completed: alter database open
Tue Sep 02 21:40:19 2008
KCF: write/open error block=0x1bf online=1
     file=2 E:\BARCODE\ORADATA\BARCODE\UNDOTBS01.DBF
     error=27072 txt: 'OSD-04008: WriteFile() 失败, 无法写入文件
O/S-Error: (OS 33) 另一个程序已锁定文件的一部分，进程无法访问。'
Automatic datafile offline due to write error on
file 2: E:\BARCODE\ORADATA\BARCODE\UNDOTBS01.DBF
Tue Sep 02 21:40:19 2008
Errors in file e:\barcode\admin\barcode\bdump\barcode_smon_3236.trc:
ORA-00376: file 2 cannot be read at this time
ORA-01110: data file 2: 'E:\BARCODE\ORADATA\BARCODE\UNDOTBS01.DBF'

我们知道，windows系统比较特殊，当一个程序打开一个文件之后，文件就不能被其他进程修改或删除(unix可以).所以分析上面的原因就是这个特性引起的。但是引起的问题的具体原因还要分析。

下面贴出我的回复，算是一个分析思路：

先转两个metalink的文档
重点看一下红字

1)   Taking hotbackup of the database in Windows environment.
The third party tool used for taking hotbackup was Veritas Backup Exec.
At the time of the backup the following messages are displayed in alert.log.

KCF: write/open error block=0x76b online=1
file=11 I:\APPS\ASCENDENT\DATABASE\ATP1ORCL.ORA
error=27072 txt: 'OSD-04008: WriteFile() failure, unable to write to file
O/S-Error: (OS 33) The process cannot access the file because another process
has locked a portion of the file.'
Automatic datafile offline due to write error on
file 11: I:\APPS\ASCENDENT\DATABASE\ATP1ORCL.ORA
Mon May 31 14:58:20 2004
Errors in file i:\oracle\admin\atp\udump\atp_ora_3564.trc:
ORA-00202: controlfile: 'I:\APPS\ASCENDENT\DATABASE\ATP\CONTROL02.CTL'
ORA-27091: skgfqio: unable to queue I/O
ORA-27070: skgfdisp: async read/write failed
OSD-04006: ReadFile() failure, unable to read from file
O/S-Error: (OS 33) The process cannot access the file because another process
has locked a portion of the file.
Cause
The backup utility (Veritas Backup Exec) has locked the file for taking the backup and Oracle is not able to access the file exclusively.

Fix
Disabling file locking option in Veritas BackupExec resolved the issue.


2）This error usually comes accompanied with error: <ORA-27072> : skgfdisp: I/O error
Which is caused by read/write/readv/writev system call returned error, additional
information indicates starting block number of I/O

The operating system is reporting the error : O/S-Error: (OS 33) The process cannot access the file because another process has locked a portion of the file.

In brief, NT-33 is an error_lock_violation indicating that a portion of the data file is locked by
another NT process.

This is NOT an Oracle problem as the data file is locked at the OS level by another process other than
Oracle.

You should also check whether there is software that accesses the database files in order to see why the files are reported as being locked (i.e. : a Virus Scanner etc.).


References / Known Bugs :
~~~~~~~~~~~~~~~~~~
There are several known issues that can cause this error, plus it can be
a non-Oracle problem so it is important to collect all the information:

Note 62427.1 :  2GB or not 2GB - file limits in Oracle
See the above article as an anchor point for 2GB issues on various platforms.

3)
Customer tries to startup the database:
Fri Jan 04 21:53:40 2008  
Errors in file d:\oracle\admin\mndbprd\bdump\mndbprd_ckpt_3808.trc:  
ORA-01171: datafile 9 going offline due to error advancing checkpoint  
ORA-01110: data file 9: 'D:\ORACLE\ORADATA\MNDBPRD\MN_DATA101.DBF'  
ORA-01115: IO error reading block from file 9 (block # 1)  
ORA-27072: skgfdisp: I/O error  
OSD-04006: ReadFile() failure, unable to read from file  
O/S-Error: (OS 33) The process cannot access the file because another process has locked a portion of the file.

CauseA tape backup software (in this case Veritas Backup Exec) had been run and it still has a lock on the Oracle data file 9 and the OS will not allow Oracle to write to it.
SolutionYou will need to reboot the Windows server to release the lock on the datafile. (这个只是可能)

楼主你对比下，你的符合哪个

Errors in file e:\barcode\admin\barcode\bdump\barcode_lgwr_3040.trc:
ORA-00345: redo log write error block 20528 count 2
ORA-00312: online log 3 thread 1: 'E:\BARCODE\ORADATA\BARCODE\REDO03.LOG'
ORA-27072: skgfdisp: I/O error
OSD-04008: WriteFile() 失败, 无法写入文件
O/S-Error: (OS 33) 另一个程序已锁定文件的一部分，进程无法访问。

看你alertlog的这一个错误不像备份软件引起的，备份软件应该不会去锁定redo Log的
还是杀毒软件或病毒的可能性比较大