RAC 之 RMAN 恢复

之前整理的RMAN有关还原的文章：

RMAN系列（五）---- RMAN还原与恢复

http://blog.csdn.net/xujinyang/article/details/6838045

RMAN备份与恢复实例

http://blog.csdn.net/xujinyang/article/details/6822620

Oracle Rman跨resetlogs版本恢复

http://blog.csdn.net/xujinyang/article/details/6830337

RAC下的RMAN讲究的是备份和还原的策略要一致。备份策略的不同，会导致备份结果的分步不同，进而影响恢复的策略和步骤。一般情况下，恢复策略和备份策略必须是对应的。如果备份策略进行了修改，那么恢复也应该采用相同的策略。之前我们的RMAN备份参考：

RAC之RMAN备份

http://blog.csdn.net/xujinyang/article/details/6837226

这里面有2个注意事项：

1.执行恢复的节点要能看到所有节点的归档文件。这个在RAC备份那篇里已经讲了方法，就是把归档文件传送到其他节点，这样就能够看到所有节点的归档文件了。

2.备份集的存放问题。如果我们放在共享设备上，那么节点可以访问所有的备份文件，那么就可以直接进行恢复操作。如果不是放在共享设备，可以通过NFS mount来挂载，或者设置2个通道。分别指向多个实例。

示例：

1.1先停止数据库

[oracle@rac1 bin]$srvctl stop db -d orcl

[oracle@rac1 bin]$crs_stat -t

NameTypeTargetStateHost

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

ora.orcl.dbapplicationOFFLINEOFFLINE

ora....oltp.cs applicationOFFLINEOFFLINE

ora....cl1.srv applicationOFFLINEOFFLINE

ora....cl2.srv applicationOFFLINEOFFLINE

ora....l1.inst applicationOFFLINEOFFLINE

ora....l2.inst applicationOFFLINEOFFLINE

ora....SM1.asm applicationONLINEONLINErac1

ora....C1.lsnr applicationONLINEONLINErac1

ora.rac1.gsdapplicationONLINEONLINErac1

ora.rac1.onsapplicationONLINEONLINErac1

ora.rac1.vipapplicationONLINEONLINErac1

ora....SM2.asm applicationONLINEONLINErac2

ora....C2.lsnr applicationONLINEONLINErac2

ora.rac2.gsdapplicationONLINEONLINErac2

ora.rac2.onsapplicationONLINEONLINErac2

ora.rac2.vipapplicationONLINEONLINErac2

1.2将节点启动到mount状态：

备份集存在在2个节点的情况（2个节点的RAC），在这种情况下，需要分配2个通道分别连接对应的节点。每个节点都需要启动到mount状态。因为每个节点都有备份文件，所以在每个节点都需要执行restore，如果没有启动的话，就会报错，对于备份集存放在共享设备的就不用这么麻烦。

[oracle@rac1 bin]$export ORACLE_SID=orcl1

[oracle@rac1 bin]$rman target /

Recovery Manager: Release 10.2.0.1.0 - Production on Mon Sep 27 02:56:55 2010

Copyright (c) 1982, 2005, Oracle.All rights reserved.

connected to target database (not started)

RMAN>startup mount;

Oracle instance started

database mounted

Total System Global Area180355072 bytes

Fixed Size1218388 bytes

Variable Size100665516 bytes

Database Buffers75497472 bytes

Redo Buffers2973696 bytes

[oracle@rac2 bin]$ export ORACLE_SID=orcl2

[oracle@rac2 bin]$ rman target /

Recovery Manager: Release 10.2.0.1.0 - Production on Mon Sep 27 02:57:44 2010

Copyright (c) 1982, 2005, Oracle.All rights reserved.

connected to target database (not started)

RMAN> startup mount;

Oracle instance started

database mounted

Total System Global Area180355072 bytes

Fixed Size1218388 bytes

Variable Size83888300 bytes

Database Buffers92274688 bytes

Redo Buffers2973696 bytes

1.3完全恢复

在任意一个节点上执行：

RMAN>RUN {

allocate channel c1 device type disk connect'sys/oracle@orcl1';

allocate channel c2 device type disk connect'sys/oracle@orcl2';

restore database;

recover database;

}

2> 3> 4> 5> 6>

using target database control file instead of recovery catalog

allocated channel: c1

channel c1: sid=146 instance=orcl1 devtype=DISK

allocated channel: c2

channel c2: sid=146 instance=orcl2 devtype=DISK

Starting restore at 27-SEP-10

skipping datafile 2; already restored to file +DATA/orcl/datafile/undotbs1.278.730181053

skipping datafile 3; already restored to file +DATA/orcl/datafile/sysaux.277.730181053

skipping datafile 5; already restored to file +DATA/orcl/datafile/undotbs2.284.730181347

skipping datafile 1; already restored to file +DATA/orcl/datafile/system.276.730181051

skipping datafile 4; already restored to file +DATA/orcl/datafile/users.279.730181053

skipping datafile 7; already restored to file +DATA/orcl/datafile/anhuianqing.dbf

Finished restore at 27-SEP-10

Starting recover at 27-SEP-10

starting media recovery

media recovery complete, elapsed time: 00:00:04

Finished recover at 27-SEP-10

released channel: c1

released channel: c2

RMAN> sql 'alter database open';

sql statement: alter database open

1.4不完全恢复

1.4.1基于时间点的恢复

RMAN>RUN {

allocate channel c1 device type disk connect'sys/oracle@orcl1';

allocate channel c2 device type disk connect'sys/oracle@orcl2';

set until time "to_date('2010/9/27 03:30:00','yyyy/mm/dd hh24:mi:ss')";

restore database;

recover database;

}

2>3> 4> 5> 6> 7> 8>

allocated channel: c1

channel c1: sid=143 instance=orcl1 devtype=DISK

allocated channel: c2

channel c2: sid=146 instance=orcl2 devtype=DISK

executing command: SET until clause

Starting restore at 27-SEP-10

channel c1: starting datafile backupset restore

channel c1: specifying datafile(s) to restore from backup set

restoring datafile 00001 to +DATA/orcl/datafile/system.276.730181051

restoring datafile 00004 to +DATA/orcl/datafile/users.279.730181053

restoring datafile 00007 to +DATA/orcl/datafile/anhuianqing.dbf

channel c1: reading from backup piece /u01/backup/orcl_2blotmnp_1_1_20100927

channel c2: starting datafile backupset restore--这里使用了2个通道，因为这个测试的备份集是放在2个节点的。

channel c2: specifying datafile(s) to restore from backup set

restoring datafile 00002 to +DATA/orcl/datafile/undotbs1.278.730181053

restoring datafile 00003 to +DATA/orcl/datafile/sysaux.277.730181053

restoring datafile 00005 to +DATA/orcl/datafile/undotbs2.284.730181347

channel c2: reading from backup piece /u01/backup/orcl_2clotmno_1_1_20100927

channel c2: restored backup piece 1

piece handle=/u01/backup/orcl_2clotmno_1_1_20100927 tag=ORCL_HOT_DB_BK

channel c2: restore complete, elapsed time: 00:01:21

channel c1: restored backup piece 1

piece handle=/u01/backup/orcl_2blotmnp_1_1_20100927 tag=ORCL_HOT_DB_BK

channel c1: restore complete, elapsed time: 00:02:48

Finished restore at 27-SEP-10

Starting recover at 27-SEP-10

starting media recovery

media recovery complete, elapsed time: 00:00:03

Finished recover at 27-SEP-10

released channel: c1

released channel: c2

RMAN>

恢复完成之后，在两个节点执行：

alter database open resetlogs;

打开数据即可。

1.4.2基于SCN的恢复

当前SCN查看方法：

SQL> select current_scn from v$database;

CURRENT_SCN

-----------

1118216

RMAN>RUN {

allocate channel c1 device type disk connect'sys/oracle@orcl1';

allocate channel c2 device type disk connect'sys/oracle@orcl2';

restore database until scn 1118215 ;

recover database until scn 1118215 ;

}

2> 3> 4> 5> 6>

using target database control file instead of recovery catalog--非catalog模式，这种模式备份控制文件非常重要

allocated channel: c1

channel c1: sid=142 instance=orcl1 devtype=DISK

allocated channel: c2

channel c2: sid=151 instance=orcl2 devtype=DISK

Starting restore at 27-SEP-10

channel c1: starting datafile backupset restore

channel c1: specifying datafile(s) to restore from backup set

restoring datafile 00001 to +DATA/orcl/datafile/system.276.730181051

restoring datafile 00004 to +DATA/orcl/datafile/users.279.730181053

restoring datafile 00007 to +DATA/orcl/datafile/anhuianqing.dbf

channel c1: reading from backup piece /u01/backup/orcl_2blotmnp_1_1_20100927

channel c2: starting datafile backupset restore

channel c2: specifying datafile(s) to restore from backup set

restoring datafile 00002 to +DATA/orcl/datafile/undotbs1.278.730181053

restoring datafile 00003 to +DATA/orcl/datafile/sysaux.277.730181053

restoring datafile 00005 to +DATA/orcl/datafile/undotbs2.284.730181347

channel c2: reading from backup piece /u01/backup/orcl_2clotmno_1_1_20100927

channel c2: restored backup piece 1

piece handle=/u01/backup/orcl_2clotmno_1_1_20100927 tag=ORCL_HOT_DB_BK

channel c2: restore complete, elapsed time: 00:03:58

channel c1: restored backup piece 1

piece handle=/u01/backup/orcl_2blotmnp_1_1_20100927 tag=ORCL_HOT_DB_BK

channel c1: restore complete, elapsed time: 00:05:47

Finished restore at 27-SEP-10

Starting recover at 27-SEP-10

starting media recovery

archive log thread 1 sequence 44 is already on disk as file /u02/rac1_arch/1_44_730181171.dbf

archive log thread 2 sequence 38 is already on disk as file /u02/rac1_arch/2_38_730181171.dbf

archive log thread 1 sequence 1 is already on disk as file /u02/rac1_arch/1_1_730784806.dbf

archive log thread 2 sequence 1 is already on disk as file /u02/rac2_arch/2_1_730784806.dbf

archive log filename=/u02/rac1_arch/1_44_730181171.dbf thread=1 sequence=44

archive log filename=/u02/rac1_arch/2_38_730181171.dbf thread=2 sequence=38

archive log filename=/u02/rac2_arch/2_1_730784806.dbf thread=2 sequence=1

media recovery complete, elapsed time: 00:00:09

Finished recover at 27-SEP-10

released channel: c1

released channel: c2

恢复结束，用alter database open resetlogs打开数据库。

小结：

以上都是简单的测试，具体情况还是要具体对待。经验积累的多了，就知道怎么处理了。总之，备份重于一切。有了备份，才有恢复的可能。

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