只读表空间的备份与恢复

一、只读表空间的特性

使用只读表空间避免对静态数据的频繁备份

当使用alter
tablespace tbs read only时,数据文件会执行检查点进程(将所有脏缓冲区的内容写至磁盘)，

当前的SCN号会被标注,同时存储了SCN的数据文件头部被冻结.控制文件内也会记录该数据文件的冻结信息。

可以清除只读表空间的对象

二、只读表空间的备份

一般情况下，只读表空间只需要进行一次备份，即当表空间状态发生改变时应立即进行备份

可以使用OS系统cp命令来备份或RMAN进行备份只读表空间

使用RMAN时建议启用备份优化选项

RMAN> CONFIGURE BACKUP OPTIMIZATION ON;

只读表空间不支持热备

SQL> alter tablespace tbs1 begin backup;

alter
tablespace tbs1 begin backup

*

ERROR at
line 1:

ORA-01642:
begin backup not needed for read only tablespace 'TBS1'

三、只读表空间的还原与恢复

还原与恢复只读表空间的问题在于控制文件如何控制只读表空间，分为下列三种情况：

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

case
backup
1
crash
status
recovery

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

case
1
Read-Only
Read-Only
将备份的只读表空间复制到目的地(Restore)

case
2
Read-Only
Read-Write
先Restore backup1,后recover(applied log )

case
3
Read-Write
Read-only
先Restore backup1,后recover(applied log )

只读表空间恢复时需要考虑的问题

重建一个控制文件时

重命名数据文件时

使用一个备份的控制文件时

下面对表空间tbs1置为只读后对比前后生成的重建控制文件的脚本

SQL> alter database backup controlfile to trace as
'/tmp/rectl1.sql';

SQL> alter tablespace tbs1 read
only;

SQL> alter database backup controlfile to trace as
'/tmp/rectl2.sql';

SQL> ho diff /tmp/rectl1.sql
/tmp/rectl2.sql

69,70c69

<
'/u01/app/oracle/oradata/orcl/example01.dbf',

<
'/u01/app/oracle/oradata/orcl/tbs01.dbf'

---

>
'/u01/app/oracle/oradata/orcl/example01.dbf'

97a97,102

> -- Files in read-only tablespaces are now
named.

> ALTER DATABASE RENAME FILE
'MISSING00006'

> TO
'/u01/app/oracle/oradata/orcl/tbs01.dbf';

>

> -- Online the files in read-only
tablespaces.

> ALTER TABLESPACE "TBS1"
ONLINE;

对比两者

1.使用create controlfile命令时，datafile中未列出只读表空间的数据文件

2.成功创建控制文件并打开后，使用alter database rename file命令重命名只读表空间的数据文件

3.使用alter tablespace readonly_tablespacename online
将只读表空间联机

四、演示只读表空间变化的恢复过程

1.
演示整个过程为只读表空间的情况(对应前面描述的case
1)

SQL> create table scott.tb1 tablespace
tbs1

2 as select * from
scott.emp;

SQL>
commit;

SQL> alter tablespace tbs1 read
only;

SQL> select file#,name,enabled from v$datafile where
file#=6;

FILE#
NAME
ENABLED

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

6
/u01/app/oracle/oradata/orcl/tbs01.dbf
READ
ONLY

SQL> ho cp /u01/app/oracle/oradata/orcl/tbs01.dbf
/tmp/tbs01.dbf

SQL> insert into scott.tb1(empno,ename)
values(3333,'Thomas');

SQL> update scott.tb1 set sal=sal*1.2 where
ename='SCOTT';

SQL> delete from scott.tb1 where
ename='SCOTT';

--执行上述三条命令，收到下列同样的错误提示

ORA-00372: file 6 cannot be modified at this
time

ORA-01110: data file 6:
'/u01/app/oracle/oradata/orcl/tbs01.dbf'

从上面的演示可以看出对只读表空间内的数据作任何DML操作均不可用

--使用vim
打开/u01/app/oracle/oradata/orcl/tbs01.dbf文件做任意操作来模拟破坏该只读表空间的数据文件

--重启数据库后收到下面的错误提示

ORA-01157: cannot identify/lock data file 6 - see DBWR trace
file

ORA-01110: data file 6:
'/u01/app/oracle/oradat/orcl/tbs01.dbf'

SQL> ho cp /tmp/tbs01.dbf
/u01/app/oracle/oradata/orcl/tbs01.dbf

SQL> alter database
open;

SQL> select count(1) from
scott.tb1;

COUNT(1)

----------

16

2.
演示由只读变为读写后发生损坏且只有只读备份的恢复情况(对应前面描述的case
2)

SQL> alter tablespace tbs1 read
write;

SQL> insert into scott.tb1(empno,ename)
values(3333,'Thomas');

SQL> commit;

--使用vim
打开/u01/app/oracle/oradata/orcl/tbs01.dbf文件做任意操作来模拟破坏该读写表空间的数据文件

--重启数据库后未收到错误提示

SQL> insert into scott.tb1(empno,ename)
values(4444,'Jackson');

insert into scott.tb1(empno,ename) values(4444,'Jackson')

*

ERROR at line 1:

ORA-00376:
file 6 cannot be read at this time

ORA-01110: data file 6:
'/u01/app/oracle/oradata/orcl/tbs01.dbf'

SQL> select * from v$recover_file;

　　FILE#
ONLINE ONLINE_
ERROR
CHANGE# TIME

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

　　6

OFFLINE OFFLINE FILE NOT
FOUND
0

--使用OS的cp命令还原，然后recover

　　SQL> ho cp /tmp/tbs01.dbf
/u01/app/oracle/oradata/orcl/tbs01.dbf

　　SQL> recover datafile 6;

　　Media recovery complete.

　　SQL> alter tablespace tbs1 online;

　　Tablespace altered.

　　SQL> select * from scott.tb1 where
ename='Thomas';

　　
EMPNO
ENAME
JOB
MGR
HIREDATE
SAL
COMM
DEPTNO

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

　　
3333 Thomas

3.
演示由读写表空间变为只读表空间，且仅有读写表空间备份的恢复(对应前面描述的case
3)

SQL> select file#,name,enabled from v$datafile where
file#=6;

FILE#
NAME
ENABLED

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

6
/u01/app/oracle/oradata/orcl/tbs01.dbf
READ
WRITE

SQL> alter tablespace tbs1 begin
backup;

SQL> ho cp /u01/app/oracle/oradata/orcl/tbs01.dbf
/tmp/tbs01.dbf

SQL> alter tablespace tbs1 end
backup;

SQL> delete from scott.tb1 where
empno=3333;

SQL>
commit;

SQL> alter tablespace tbs1 read
only;

--使用vim
打开/u01/app/oracle/oradata/orcl/tbs01.dbf文件做任意操作来模拟破坏该只读表空间的数据文件

--重启数据库后收到下面的错误提示

ORA-01157: cannot identify/lock data file 6 - see DBWR trace
file

ORA-01110: data file 6:
'/u01/app/oracle/oradata/orcl/tbs01.dbf'

SQL> ho cp /tmp/tbs01.dbf
/u01/app/oracle/oradata/orcl/tbs01.dbf

SQL> recover datafile
6;

SQL> alter database
open;

SQL> select * from scott.tb1 where
ename='Thomas';

no rows
selected

4.
演示数据文件状态发生多次变化且在变化时没有任何备份的恢复处理(实际上使用日志来重新构造该数据文件)

SQL> select file#,name,enabled from v$datafile where
file#=6;

FILE#
NAME
ENABLED

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

6
/u01/app/oracle/oradata/orcl/tbs01.dbf
READ
WRITE

SQL> drop table
scott.tb1;

SQL>
commit;

SQL> alter tablespace tbs1 read
only;

SQL> alter tablespace tbs1 read
write;

SQL> create table scott.tb2 tablespace tbs1 as
select * from
scott.emp;

SQL>
commit;

SQL> alter system
checkpoint;

[oracle@oradb orcl]$tail -n 50
$ORACLE_BASE/admin/orcl/bdump/alert_orcl.log

Errors in file
/u01/app/oracle/admin/orcl/bdump/orcl_ckpt_4064.trc:

ORA-01171: datafile 6 going offline due to error advancing
checkpoint

ORA-01122: database file 6 failed verification
check

ORA-01110: data file 6:
'/u01/app/oracle/oradata/orcl/tbs01.dbf'

ORA-01251: Unknown File Header Version read for file number
6

SQL> select * from
v$recover_file;

FILE# ONLINE ONLINE_
ERROR
CHANGE#
TIME

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

6 OFFLINE OFFLINE FILE NOT
FOUND
0

SQL> select file#,name,status from v$datafile where
file#=6;

FILE#
NAME
STATUS

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

6
/u01/app/oracle/oradata/orcl/tbs01.dbf
RECOVER

SQL> alter database create datafile
6;

SQL> recover datafile
6;

Media recovery
complete.

SQL> alter database datafile 6
online;

SQL> select count(1) from
scott.tb2;

COUNT(1)

----------

16

5.
演示删除只读表空间上的对象

SQL> select file#,name,enabled from v$datafile where
file#=6;

FILE#
NAME
ENABLED

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

6
/u01/app/oracle/oradata/orcl/tbs01.dbf
READ
ONLY

SQL> select
segment_name,segment_type,tablespace_name,owner from dba_segments
where

2 tablespace_name='TBS1' and
segment_name='TB2';

SEGMENT_NAME
SEGMENT_TYPE
TABLESPACE_NAME
OWNER

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

TB2
TABLE
TBS1
SCOTT

SQL> drop table
scott.tb2;

Table
dropped.

五、总结

1.
表空间置为只读后将减少数据的备份量

2.
表空间置为只读后，不能对其中的对象执行任何DML操作

3.
只读表空间内的对象可以被清除，因为drop命令更新了数据字典，而不更新对象本身

4.
当表空间的状态发生变化时，应立即备份该表空间，以减少恢复工作

5.
对于状态多次发生改变且未及时备份的情况，日志未损坏时，可以使用联机重做、归档日志来进行恢复

使用下列命令来实现：

删除受损的数据文件(rm dbfile.dbf)

重建受损的数据文件(alter database create datafile n)

进行介质恢复(recover datafile n)

使受损的数据文件联机(alter database datafile n online)

6.
演示中多为在mount状态下来恢复，生产环境中多在open状态下恢复，可以按下列步骤实现

先将受损的只读表空间(数据文件)脱机(offline)

使用备份的表空间(数据文件)来还原(restore)

使用归档、联机日志进行介质恢复(recover)

使恢复成功的表空间(数据文件)联机(online)

7.
对于原始介质受损，不能恢复到原始位置的情况下,使用下面的命令实现转移

alter database rename file '' to '';

转自：http://blog.csdn.net/robinson_0612/article/details/6646492

另附OCP试题

153. You have a read­only tablespace on read­only media. You want to perform a media recovery on the

existing data files, but using a backup control file. The backup control file indicates that the status of the

read­only tablespace was read/write when the control file was backed up.

What should you consider?

A.drop the read­only tablespace and re­create the tablespace after recovery

B.take data files from the read­only tablespace offline before performing a recovery

C.perform a recovery; status of the tablespace in control file will be changed automatically

D.recovery using backup control file is not possible, so restore all the files from the last full database

backup, and then open the database

Answer: B

154. The loss of the control file forced you to re­create the control file. After re­creating it and opening the

database, you find that some of the data files are named as MISSINGnnnnn, where nnnnn is a five­digit

number starting with 0.

What could be the possible reason?

A.These are the data files that are corrupted.

B.There is no SYSAUX tablespace in your database

C.There is a read­only tablespace in your database.

D.These are the data files that cannot be recovered.

Answer: C