使用RMAN异机恢复磁带库的备份文件至文件系统

描述：原生产环境为Linux RAC+ASM+10.2.0.4，因存储故障，数据库无法启动；需使用第三方磁带库备份软件备份的文件进行数据库恢复。

恢复控制文件:

SET DBID 2240547690;
run {
allocate channel ch1 type 'sbt_tape' PARMS="SBT_LIBRARY=/opt/simpana/Base/libobk.so";
restore controlfile to '/home/oracle/control01.ctl'  from autobackup ;
release channel ch1;
}


启动数据库至mount状态：
alter database mount;


restore database：
run {
allocate channel ch1 type 'sbt_tape' PARMS="SBT_LIBRARY=/opt/simpana/Base/libobk.so";
set newname for datafile 1  to '/oradata/orcl/system.256.769887829 ';
set newname for datafile 2  to '/oradata/orcl/undotbs1.258.769887829 ';
set newname for datafile 3  to '/oradata/orcl/sysaux.257.769887829 ';
set newname for datafile 4  to '/oradata/orcl/users.259.769887829 ';
set newname for datafile 5  to '/oradata/orcl/undotbs2.264.769887929 ';
set newname for datafile 6  to '/oradata/orcl/undotbs3.265.769887929 ';
set newname for datafile 7  to '/oradata/orcl/user01 ';
set newname for datafile 8  to '/oradata/orcl/user02 ';
restore database; 
switch datafile all ; 
release channel ch1;
}


recover database：
run {
allocate channel ch1 type 'sbt_tape' PARMS="SBT_LIBRARY=/opt/simpana/Base/libobk.so";
recover database;
release channel ch1;
}


更改redolog文件目录：
ALTER DATABASE RENAME FILE '+DATA01/orcl/onlinelog/group_2.262.769887895'  TO '/oradata/orcl/group_2.262.769887895';
ALTER DATABASE RENAME FILE '+DATA01/orcl/onlinelog/group_1.261.769887889'  TO '/oradata/orcl/group_1.261.769887889';
ALTER DATABASE RENAME FILE '+DATA01/orcl/onlinelog/group_5.266.769887971'  TO '/oradata/orcl/group_5.266.769887971';
ALTER DATABASE RENAME FILE '+DATA01/orcl/onlinelog/group_6.267.769887977'  TO '/oradata/orcl/group_6.267.769887977';
ALTER DATABASE RENAME FILE '+DATA01/orcl/onlinelog/group_3.268.769887983'  TO '/oradata/orcl/group_3.268.769887983';
ALTER DATABASE RENAME FILE '+DATA01/orcl/onlinelog/group_4.269.769887989'  TO '/oradata/orcl/group_4.269.769887989';


启动数据库至OPEN状态：
SQL> alter database open resetlogs;
Database altered.


删除原有的临时表空间文件：
alter database tempfile '+DATA01/orcl/tempfile/temp.263.769887907'  offline;
alter database tempfile '+DATA01/orcl/tempfile/temp.263.769887907'  drop;


添加新的临时表空间文件：

alter tablespace  TEMP            add tempfile '/oradata/orcl/temp02'    size   7636M autoextend on  maxsize unlimited ;

参考MOS文章如下：

How To Restore Rman Backups On A Different Node When The Directory Structures Are Different (文档 ID 419137.1)

In this Document

Goal

Solution

Steps to acheive the goal:

1) Connect to the target database using rman and backup the database ---> ON NODE 1

3) Edit the PFILE on NODE 2 to change the environment specific parameters like .

4) Once the PFILE is suitably modified invoke Rman on the NODE 2 after setting the Oracle environment variables and start the database in nomount mode:  

5) Restore the controlfile from the backup piece.

7) Now catalog the backup pieces that were shipped from NODE 1 

8) Get to know the last sequence available in the archivelog backup using the following command.This will help us in recovering the database till that archivelog.

9) Rename the Redolog files,so that they can be created in new locations when opened the database is opened in resetlogs 

10) Now restore the datafiles to new locations and recover. Since we are recovering the database here till the archivelog sequence 50 the sequence number in the SET UNTIL SEQUENCE clause

11) Confirm your tempfiles

References

---

This document is being delivered to you via Oracle Support's Rapid Visibility (RaV) process and therefore has not been subject to an independent technical review.

APPLIES TO:

Oracle Database - Enterprise Edition - Version 10.1.0.2 and later
Information in this document applies to any platform.
***Checked for relevance on 20-July-2015*** 

GOAL

Restoring Rman backup on a different node with different backup directory structures and different database directory structures . 

++ Rman backup of database is done on NODE 1. 
++ The database has to be has to be restored from the Rman backup on NODE 2. 
++ The directory structures for the backups and the database files are different on NODE 2. 
++ The Rman backups have to put in new directory structure on NODE 2 unlike as they were on NODE 1. 
++ Also the database has to be restored to a different directory structure on NODE 2. 

SOLUTION

ASSUMPTIONS 
-----------------------

++ The backups on NODE 1 are done to '/node1/database/backup' 
++ The database files on NODE 1 are in '/node1/database/prod' 
++ The backups on NODE 2 will be in location '/node2/database/backup' 
++ The database files on NODE 2 will be restored to '/node2/database/prod

Steps to acheive the goal:

1) Connect to the target database using rman and backup the database ---> ON NODE 1

$ rman target / 

Recovery Manager: Release 10.2.0.1.0 - Production on Tue Feb 13 00:29:33 2007 
Copyright (c) 1982, 2005, Oracle. All rights reserved. 
connected to target database: ORA10G (DBID=3932056136) 


RMAN> backup database plus archivelog; 


Starting backup at 13-FEB-07 
current log archived 
using channel ORA_DISK_1 
channel ORA_DISK_1: starting archive log backupset 
channel ORA_DISK_1: specifying archive log(s) in backup set 
input archive log thread=1 sequence=143 recid=109 stamp=614392105 
channel ORA_DISK_1: starting piece 1 at 13-FEB-07 
channel ORA_DISK_1: finished piece 1 at 13-FEB-07 
piece handle=/node1/database/backup/o1_mf_annnn_TAG20070213T002825_2x21kbds 
_.bkp tag=TAG20070213T002825 comment=NONE 
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:02 
Finished backup at 13-FEB-07 

Starting backup at 13-FEB-07 
using channel ORA_DISK_1 
channel ORA_DISK_1: starting full datafile backupset 
channel ORA_DISK_1: specifying datafile(s) in backupset 
input datafile fno=00003 name=/node1/database/prod/sysaux01.dbf 
input datafile fno=00001 name=/node1/database/prod/system01.dbf 
input datafile fno=00002 name=/node1/database/prod/undotbs01.dbf 
input datafile fno=00004 name=/node1/database/prod/users01.dbf 
input datafile fno=00005 name=/node1/database/prod/1.dbf 
input datafile fno=00006 name=/node1/database/prod/sysaux02.dbf 
input datafile fno=00007 name=/node1/database/prod/undotbs02.dbf 
channel ORA_DISK_1: starting piece 1 at 13-FEB-07 
channel ORA_DISK_1: finished piece 1 at 13-FEB-07 
piece handle=/node1/database/backup/o1_mf_nnndf_TAG20070213T002827_2x21kd12 
_.bkp tag=TAG20070213T002827 comment=NONE 
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:55 
Finished backup at 13-FEB-07 


Starting backup at 13-FEB-07 
current log archived 
using channel ORA_DISK_1 
channel ORA_DISK_1: starting archive log backupset 
channel ORA_DISK_1: specifying archive log(s) in backup set 
input archive log thread=1 sequence=144 recid=110 stamp=614392165 
channel ORA_DISK_1: starting piece 1 at 13-FEB-07 
channel ORA_DISK_1: finished piece 1 at 13-FEB-07 
piece handle=/node1/database/backup/o1_mf_annnn_TAG20070213T002925_2x21m6ty 
_.bkp tag=TAG20070213T002925 comment=NONE 
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:02 
Finished backup at 13-FEB-07 


Starting Control File and SPFILE Autobackup at 13-FEB-07 
piece handle=/u01/oracle/product/ora10g/dbs/c-3932056136-20070213-02 comment=NONE 
Finished Control File and SPFILE Autobackup at 13-FEB-07 


RMAN> exit

2) Move the following files to the NODE 2:

+ The database backup pieces to location '/node2/database/backup' 
+ Controlfile backup piece to the location '/node2/database/backup' 
+ The parameter file i.e init.ora file to the default location i.e $ORACLE_HOME/dbs

3) Edit the PFILE on NODE 2 to change the environment specific parameters like .

user_dump_dest =  
background_dump_dest = 
control_files =

4) Once the PFILE is suitably modified invoke Rman on the NODE 2 after setting the Oracle environment variables and start the database in nomount mode:  

[oracle@test-br test]$ export ORACLE_HOME=/u01/oracle/product/ora10g 
[oracle@test-br test]$ export ORACLE_SID=ora10g 
[oracle@test-br test]$ export PATH=$ORACLE_HOME/bin:$PATH 
[oracle@test-br test]$ rman target / 
Recovery Manager: Release 10.2.0.1.0 - Production on Tue Feb 13 00:36:55 2007 
Copyright (c) 1982, 2005, Oracle.  All rights reserved. 
connected to target database (not started) 
RMAN> startup nomount 
Oracle instance started 
Total System Global Area     205520896 bytes 
Fixed Size                     1218508 bytes 
Variable Size                 75499572 bytes 
Database Buffers             121634816 bytes 
Redo Buffers                   7168000 bytes

5) Restore the controlfile from the backup piece.

RMAN> restore controlfile from  '/node2/database/backup/c-3932056136-20070213-02';  
Starting restore at 13-FEB-07  
using target database control file instead of recovery catalog  
allocated channel: ORA_DISK_1  
channel ORA_DISK_1: sid=155 devtype=DISK  
channel ORA_DISK_1: restoring control file  
channel ORA_DISK_1: restore complete,  
elapsed time: 00:00:02  
output filename=/node2/database/prod/control01.ctl  
Finished restore at 13-FEB-07 

6) Mount the database

RMAN > alter database mount      
     

7) Now catalog the backup pieces that were shipped from NODE 1 

RMAN> catalog backuppiece '/node2/database/backup/o1_mf_annnn_TAG20070213T002925_2x21m6ty_.bkp'; 

RMAN> catalog backuppiece '/node2/database/backup/o1_mf_annnn_TAG20070213T002825_2x21kbds_.bkp'; 

RMAN> catalog backuppiece '/node2/database/backup/o1_mf_nnndf_TAG20070213T002827_2x21kd12_.bkp'; 
 

 

This feature of cataloging backup pieces is available from ORACLE 10g versions. Prior to Oracle 10g we were not able to catalog the backup pieces. For more information on cataloging options refer the metalink note 470463.1                 

8) Get to know the last sequence available in the archivelog backup using the following command.This will help us in recovering the database till that archivelog.

 

RMAN > list backup of archivelog all;

Let us assume the last sequence of last archivelog in the backup is 50.

9) Rename the Redolog files,so that they can be created in new locations when opened the database is opened in resetlogs 

SQL> alter database rename file '/node1/database/prod/redo01.log' to '/node2/database/prod/redo01.log'; ...... ...... ...... 

10) Now restore the datafiles to new locations and recover. Since we are recovering the database here till the archivelog sequence 50 the sequence number in the SET UNTIL SEQUENCE clause

RMAN> run  {  set until sequence 51;  set newname for datafile 1 to '/node2/database/prod/sys01.dbf';  set newname for datafile 2 to '/node2/database/prod/undotbs01.dbf';  set newname for datafile 3 to '/node2/database/prod/sysaux01.dbf';  set newname for datafile 4 to '/node2/database/prod/users01.dbf';  set newname for datafile 5 to '/node2/database/prod/1.dbf';  set newname for datafile 6 to '/node2/database/prod/sysaux02.dbf';  set newname for datafile 7 to '/node2/database/prod/undotbs02.dbf';  restore database;  switch datafile all;  recover database;  alter database open resetlogs;  } 

 

If we are restoring the Rman backups  from tapes,then we should ensure the same media manager variables that were used during backups are maintained during restore too. 

The below article gives the information of various media manager Environment Variables. 

NOTE.312737.1 RMAN and Specific Media Managers Environment Variables.

11) Confirm your tempfiles

Confirm the location of your tempfiles, recreate them at the new location if required:

SQL> select * from v$tempfile;
SQL> ALTER TABLESPACE TEMP ADD TEMPFILE 'new tempfile path filename' REUSE;