如何将RAC数据库的 RMAN Disk 备份 Restore 到另一个节点上的单个实例 (Doc ID 415579.1)
HowTo Restore RMAN Disk backups of RAC Database to Single Instance On Another Node (Doc ID 415579.1)
APPLIES TO:
Oracle Database - Enterprise Edition - Version 9.2.0.1 and later
Oracle Database Cloud Schema Service - Version N/A and later
Oracle Database Exadata Cloud Machine - Version N/A and later
Oracle Cloud Infrastructure - Database Service - Version N/A and later
Oracle Database Backup Service - Version N/A and later
Information in this document applies to any platform.
***Checked for relevance on 15-Apr-2014***
GOAL
- You have a RAC database backed up by RMAN to disk location 您有一个由RMAN备份到磁盘位置的RAC数据库
- You need to restore this backup as Single Instance on another node 您需要将此备份还原为另一个节点上的单实例
SOLUTION
1) Take appropriate RMAN backup of the production RAC database. Note that you should turn on the CONTROLFILE AUTOBACKUP configuration so that we have the controlfile backed up after the database backup. When we restore the controlfile on new host from this autobackup piece, it will have the information of the latest backup.
对生产RAC数据库进行适当的RMAN备份。请注意,您应该打开 CONTROLFILE AUTOBACKUP 配置,以便在数据库备份之后备份控制文件。当我们从此自动备份片段将控制文件还原到新主机上时,它将具有最新备份的信息。
RMAN> run{ 2> allocate channel c1 type disk format '/<PATH>/%U'; 3> backup database; 4> backup archivelog all; 5> } allocated channel: c1 channel c1: sid=133 instance=<oracle_sid> devtype=DISK Starting backup at 12-FEB-07 channel c1: starting full datafile backupset channel c1: specifying datafile(s) in backupset input datafile fno=00001 name=/<PATH>/system01.dbf input datafile fno=00002 name=/<PATH>/undotbs01.dbf input datafile fno=00005 name=/<PATH>/undotbs02.dbf input datafile fno=00003 name=/<PATH>/sysaux01.dbf input datafile fno=00004 name=/<PATH>/users01.dbf channel c1: starting piece 1 at 12-FEB-07 channel c1: finished piece 1 at 12-FEB-07 piece handle=/<PATH>/09i9sruq_1_1 tag=TAG20070212T162458 comment=NONE channel c1: backup set complete, elapsed time: 00:01:05 Finished backup at 12-FEB-07 Starting backup at 12-FEB-07 current log archived channel c1: starting archive log backupset channel c1: specifying archive log(s) in backup set input archive log thread=1 sequence=56 recid=3 stamp=613417105 input archive log thread=1 sequence=57 recid=4 stamp=613417106 input archive log thread=1 sequence=58 recid=5 stamp=614363168 input archive log thread=2 sequence=1 recid=1 stamp=613417090 input archive log thread=2 sequence=2 recid=2 stamp=613417093 input archive log thread=2 sequence=3 recid=6 stamp=614363170 channel c1: starting piece 1 at 12-FEB-07 channel c1: finished piece 1 at 12-FEB-07 piece handle=/<PATH>/0ai9ss14_1_1 tag=TAG20070212T162610 comment=NONE channel c1: backup set complete, elapsed time: 00:00:04 Finished backup at 12-FEB-07 Starting Control File and SPFILE Autobackup at 12-FEB-07 piece handle=/<PATH>/c-610677177-20070212-00 comment=NONE Finished Control File and SPFILE Autobackup at 12-FEB-07 released channel: c1
2) Create a PFILE for the single instance database using the production RAC parameter file 使用生产RAC参数文件为单实例数据库创建PFILE
a) don't forget to modify the following parameters depending on the directory structure of the new host: audit_file_dest, background_dump_dest, control_files, core_dump_dest, log_archive_dest_1, user_dump_dest etc 不要忘记根据新主机的目录结构修改以下参数:audit_file_dest, background_dump_dest, control_files, core_dump_dest, log_archive_dest_1, user_dump_dest 等
b) remove RAC specific parameters such as cluster_database_instances, cluster_database etc 删除RAC特定的参数,例如 cluster_database_instances,cluster_database 等
c) for the parameter undo_tablespace, mention any one undo tablespace name 对于参数undo_tablespace,使用任何一个undo表空间名称
Known issue which may require an additional parameter to be set : --可能需要设置其他参数的已知问题
Note 334899.1 RMAN Duplicate from RAC backup fails ORA-38856
3) Move the backup pieces and the modified INIT.ORA file to the new host. Starting from 10g it is NO longer compulsory to copy the RMAN backup pieces to exactly the same location on the new host as the production location.
将备份片段和修改后的INIT.ORA文件移动到新主机。从10g开始,不再必须将RMAN备份片段复制到新主机上与生产位置完全相同的位置。
4) Use the pfile created above to STARTUP NOMOUNT the database on the new host 使用上面创建的pfile在新主机上启动NOMOUNT数据库
oracle@test-br ractest]$ sqlplus "/ as sysdba" SQL*Plus: Release 10.2.0.1.0 - Production on Fri Feb 16 03:14:23 2007 Copyright (c) 1982, 2005, Oracle. All rights reserved. Connected to an idle instance. SQL> startup nomount; ORACLE instance started. Total System Global Area 167772160 bytes Fixed Size 1218292 bytes Variable Size 62916876 bytes Database Buffers 96468992 bytes Redo Buffers 7168000 bytes
5) Now invoke RMAN and restore the controlfile specifying the location where the controlfile autobackup piece is restored on this new server. You can mount the database once the controlfile is restored successfully.
现在调用RMAN并还原控制文件,并指定该文件在该新服务器上的还原位置。控制文件成功还原后,您可以装载数据库
[oracle@test-br ractest]$ rman target / nocatalog Recovery Manager: Release 10.2.0.1.0 - Production on Fri Feb 16 03:16:31 2007 Copyright (c) 1982, 2005, Oracle. All rights reserved. connected to target database: <db_name> (not mounted) using target database control file instead of recovery catalog RMAN> restore controlfile from '/<PATH>/c-610677177-20070212-00'; Starting restore at 16-FEB-07 allocated channel: ORA_DISK_1 channel ORA_DISK_1: sid=156 devtype=DISK channel ORA_DISK_1: restoring control file channel ORA_DISK_1: restore complete, elapsed time: 00:00:02 output filename=/<PATH>/control01.ctl Finished restore at 16-FEB-07 RMAN> alter database mount; database mounted released channel: ORA_DISK_1
6) You can skip this step if you have restored the RMAN backup pieces to exactly the same location they were backed up on production. If this is not the case then you need to catalog the RMAN backup pieces to make RMAN aware of their new location on the new host. Note that CATALOG BACKUPPIECE command is available only starting from 10g.
如果已将RMAN备份片段还原到生产中备份的完全相同的位置,则可以跳过此步骤。如果不是这种情况,则需要对RMAN备份块进行分类,以使RMAN知道它们在新主机上的新位置。注意,CATALOG BACKUPPIECE 命令仅从10g开始可用。
RMAN> catalog backuppiece '/<PATH>/09i9sruq_1_1'; cataloged backuppiece backup piece handle=/<PATH>/09i9sruq_1_1 recid=10 stamp=614661579 RMAN> catalog backuppiece '/<PATH>/0ai9ss14_1_1'; cataloged backuppiece backup piece handle=/<PATH>/0ai9ss14_1_1 recid=11 stamp=614661599
7) Now we'll determine the point upto which media recovery should run on the restored database.
现在,我们确定在恢复的数据库上应该运行介质恢复的时间点
RMAN> list backup of archivelog all; List of Archived Logs in backup set 9 Thrd Seq Low SCN Low Time Next SCN Next Time ---- ------- ---------- --------- ---------- --------- 1 56 214541 01-FEB-07 226238 01-FEB-07 1 57 226238 01-FEB-07 226240 01-FEB-07 1 58 226240 01-FEB-07 233107 12-FEB-07 2 1 186185 28-JAN-07 225714 01-FEB-07 2 2 225714 01-FEB-07 226037 01-FEB-07 2 3 226037 01-FEB-07 233110 12-FEB-07
Check the last archive sequence for all redo threads and select the archive sequence having LEAST "Next SCN" among them. In our case sequence 58 of thread 1 has Next SCN of 233107 while sequence 3 of thread 2 has Next SCN of 233110. Since squence 58 of thread 1 has least Next SCN we will recover upto this point. (If you are keen to have recovery run until some specific time you can always give SET UNTIL TIME)
检查所有 redo 线程的最后一个归档序列,并选择其中具有最低 "Next SCN" 的归档序列。在我们的情况下,线程1的序列58的Next SCN为233107,而线程2的序列3的Next SCN为233110。由于线程1的序列58的Next SCN最少,我们将恢复到这一点。(如果您希望在特定时间之前执行恢复,则可以始终设置 SET UNTIL TIME)
8) Having determined the point upto which media recovery should run, start the restore/recovery using: 确定应该运行介质恢复的时间点后,使用以下命令开始 restore/recovery
Please note in below example we choose Until sequence as 59(one more then the value selected in step 7 that is sequence 58) because When we specify a sequence number, it is considered as an UPPER limit, RMAN selects only
请使用以下步骤开始恢复:请注意,在以下示例中,我们选择 Until sequence 为59(比在步骤7中选择的值(序列58多一个)),因为当我们指定序列号时,它被视为 UPPER limit ,RMAN仅选择
those files which are needed for the recovery and uses 59 as an upper log sequence number while recovering database and would recovery upt log sequence number upto 58 .
恢复所需的文件,在恢复数据库时使用59作为较高的日志序列号,并且最多恢复58个日志序列号。
RMAN> run { 2> set until sequence 59 thread 1; 3> set newname for datafile 1 to '/<PATH>/system01.dbf'; 4> set newname for datafile 2 to '/<PATH>/undotbs01.dbf'; 5> set newname for datafile 3 to '/<PATH>/sysaux01.dbf'; 6> set newname for datafile 4 to '/<PATH>/users01.dbf'; 7> set newname for datafile 5 to '/<PATH>/undotbs02.dbf'; 8> restore database; 9> switch datafile all; 10> recover database; 11> }
set newname for datafile 1 to '+<disk_group>'; set newname for datafile 2 to '+<disk_group>'; ....
Known issue in Oracle10g:
Note 1146703.1 Rman Restore From RAC ASM To Single Instance Non ASM Fails With ORA-00600 [kgeade_is_0]
If Online Redolog's (ORL's) are Oracle Managed Files (OM), and the Clone is being done on any of Source RAC Node (where ORLs are accessible from), renaming ORL will try to remove Source ORL files!
Please review: http://docs.oracle.com/cd/E11882_01/server.112/e25494/omf.htm#ADMIN11517
The following statements are used to rename files: 以下语句用于重命名文件
ALTER DATABASE RENAME FILE ALTER TABLESPACE ... RENAME DATAFILE
These statements do not actually rename the files on the operating system, but rather, the names in the control file are changed. If the old file is an Oracle managed file and it exists, then it is deleted. You must specify each filename using the conventions for filenames on your operating system when you issue this statement."
这些语句实际不重命名操作系统上的文件,而是更改控制文件中的名称。如果旧文件是Oracle managed file并且存在,则将其删除。发出此语句时,必须使用操作系统上文件名的conventions指定每个文件名Instead of renaming, we can recreate the controlfile with desired name for ORLs: 除了重命名,我们还可以使用所需的ORL名称重新创建控制文件
SQL> alter database backup controlfile to trace as '/tmp/control.sql' resetlogs ;
Since we determined previously that media recovery should run until sequence 58 hence we use SET UNTIL SEQUENCE 59 (+1) above. You also need to use SET NEWNAME clause to restore datafiles to a location on the new host which is different from the production path. Finally, SWITCH DATAFILE ALL clause updates these new datafile locations in the controlfile.
由于我们之前确定介质恢复应运行到序列58,因此我们使用上面的 SET UNTIL SEQUENCE 59(+1)。您还需要使用SET NEWNAME子句将数据文件还原到新主机上与生产路径不同的位置。最后,SWITCH DATAFILE ALL子句更新控制文件中的这些新数据文件位置。
9) Once RMAN restore/recovery finishes, you will want to rename the online redolog files before opening the database in case the production path of redo log files is not available on the new host. After renaming the redolog files, the database can be opened with RESETLOGS
RMAN restore/recovery 完成后,如果 online redolog files 的生产路径在新主机上不可用,您将需要在打开数据库之前重命名 online redolog files。重命名重做日志文件后,可以使用RESETLOGS打开数据库
SQL> select member from v$logfile; MEMBER -------------------------------------------------------------------------------- /<old_path>/redo01.log /<old_path>/redo02.log /<old_path>/redo03.log /<old_path>/redo04.log /<old_path>/redo05.log /<old_path>/redo06.log 6 rows selected. SQL> alter database rename file '/<old_path>/redo01.log' to '/<new_path>/redo01.log'; Database altered. SQL> alter database rename file '/<old_path>/redo02.log' to '/<new_path>/redo02.log'; Database altered. SQL> alter database rename file '/<old_path>/redo03.log' to '/<new_path>/redo03.log'; Database altered. SQL> alter database rename file '/<old_path>/redo04.log' to '/<new_path>/redo04.log'; Database altered. SQL> alter database rename file '/<old_path>/redo05.log' to '/<new_path>/redo05.log'; Database altered. SQL> alter database rename file '/<old_path>/redo06.log' to '/<new_path>/redo06.log'; Database altered. SQL> alter database open resetlogs; Database altered.
如果要在ASM磁盘组(例如+ DATA)上为克隆/目标数据库创建联机重做日志,则可以使用以下命令
alter database rename file '/<old_path>/redo01.log' to '+<disk_group>'; alter database rename file '/<old_path>/redo02.log' to '+<disk_group>'; ...
10) Once the database is opened successfully, you may remove the redolog groups for redo threads of other instances
成功打开数据库后,您可以删除其他实例的重做线程的重做组
SQL> select THREAD#, STATUS, ENABLED 2 from v$thread; THREAD# STATUS ENABLED ---------- ------ -------- 1 OPEN PUBLIC 2 CLOSED PRIVATE SQL> select group# from v$log where THREAD#=2; GROUP# ---------- 4 5 6 SQL> alter database disable thread 2; Database altered. SQL> alter database drop logfile group 4; alter database drop logfile group 4 * ERROR at line 1: ORA-00350: log 4 of instance racdb2 (thread 2) needs to be archived ORA-00312: online log 4 thread 2: '/<PATH>/redo04.log' SQL> alter database clear unarchived logfile group 4; Database altered. SQL> alter database drop logfile group 4; Database altered. SQL> alter database drop logfile group 5; Database altered. SQL> alter database drop logfile group 6; Database altered. SQL> select THREAD#, STATUS, ENABLED from v$thread; THREAD# STATUS ENABLED ---------- ------ -------- 1 OPEN PUBLIC
11) Now you can remove the undo tablespaces of other instances and create a new temporary tablespace to complete the activity.
现在,您可以删除其他实例的UNDO表空间,并创建一个新的TEMP表空间以完成活动
SQL> sho parameter undo; NAME TYPE VALUE ------------------------------------ ----------- ------------------------------ undo_management string AUTO undo_retention integer 900 undo_tablespace string UNDOTBS1 SQL> SQL> SQL> select tablespace_name from dba_tablespaces where contents='UNDO'; TABLESPACE_NAME ------------------------------ UNDOTBS1 UNDOTBS2 SQL> drop tablespace UNDOTBS2 including contents and datafiles; Tablespace dropped. SQL> select name from v$tempfile; NAME -------------------------------------------------------------------------------- /<PATH>/temp01.dbf SQL> select tablespace_name from dba_tablespaces where contents='TEMPORARY'; TABLESPACE_NAME ------------------------------ TEMP SQL> create temporary tablespace TEMP1 2 tempfile '/<PATH>/temp01.dbf' 3 size 50M; Tablespace created. SQL> alter database default temporary tablespace TEMP1; Database altered. SQL> drop tablespace TEMP including contents and datafiles; Tablespace dropped.
REFERENCES
NOTE:1146703.1 - Rman Restore From RAC ASM To Single Instance Non ASM Fails With ORA-00600 [kgeade_is_0]