RAC RMAN Restore to singe node

RAC RMAN Restore to singe node

RAC RMAN RESTORE TO SINGE NODE

1. 在RAC 数据库中任意节点执行RMAN 备份脚本：

RMAN>run{

show all;

allocate channel t1 type disk connect sys/sys@mydb1;

allocate channel t2 type disk connect sys/sys@mydb2;

backup format '/home/oracle/rmandata/full_%d_%T_%U' database include current controlfile; sql 'alter system archive log current';

crosscheck archivelog all;

backup format '/Data/Backupdata/arch_%d_%T_%U' archivelog all;

release channel t1;

release channel t2;

}

备份完成之后分别在2个节点中查看备份信息：

Mydb1:

Mydb2:

RMAN 备份完成之后，在mydb1上查看tt.t1的数据及创建表tt.t_0505，分别在mydb1和mydb2上插入数据

RAC RMAN Restore to singe node

2. 在单机上安装Oracle 软件、创建监听，不创建数据库.

3. 恢复SPFILE或者PFILE，创建audit_file_dest.

?手工创建PFILE

1. 在主库上创建PFILE

2. 修改PFILE内容，删除RAC的参数，以及CONTROLFILE的路径.

?从备份中恢复SPFILE

4. Startup database nomount

5. RESTORE CONTROLFILE

RAC RMAN Restore to singe node

6. Alter Database mount，并向CONTROLFILE注册备份机，catalog start 建议'/'结尾，有些版本会因为这个找不到路径.

RAC RMAN Restore to singe node

7. 修改REDOLOG路径以及Restore database

RAC RMAN Restore to singe node

RAC RMAN Restore to singe node

........................

RAC RMAN Restore to singe node

RAC RMAN Restore to singe node

RAC RMAN Restore to singe node

8. BACKUP INCREMENT ARCHIVELOG.（全备之后的Archivelog）

9. 向CONTROLFILE注册新的ARCHIVELOG 的备份集

RAC RMAN Restore to singe node

10. RESTORE ARCHIVELOG (这步可以省略,直接跳转下一步)

RAC RMAN Restore to singe node

RAC RMAN Restore to singe node

RAC RMAN Restore to singe node

11. RECOVER DA TABASE

RAC RMAN Restore to singe node

后面的错误信息，因为Thread 1 with sequence 174 还没有产生..

12. OPEN DATABASE RESETLOGS;

RAC RMAN Restore to singe node

13. 验证数据，tt.t_0505是在FULL BACKUP 之后创建的

RAC RMAN Restore to singe node

14. Remove redolog group for redo threads of other instance

RAC RMAN Restore to singe node

15. Remove the undo tablespaces of other instances

16. Create a new temporary tablespace

RAC RMAN Restore to singe node

后续：跨版本升级RMAN..

1. RMAN RECOVR 完整之

SQL> startup upgrade; or

SQL> alter database open resetlogs upgrade;

SQL> select status from v$intsnace;

STATUS

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

OPEN MIGRATE

2. 在UPGRADE 模式下运行脚本：catupgrd.sql

SQL> spool upgrade.log

SQL>@?/rdbms/admin/catupgrd.sql

脚本执行完成，会自动shutdown database

执行此脚本，建议redo>=100M , 以及temp表空间最好自动扩展

oracle 10g 升级打oracle 11g 注意事项：

1. oracle 10g 必须是10.

2.0.2之后的版本

2. 必须在10g 库上执行@?/rdbms/admin/utlu112i.sql脚本(在11g数据库上copy)，然后在备份，否则restore之后升级时会失败，提示

SELECT TO_NUMBER('MUST_HA VE_RUN_PRE-UPGRADE_TOOL_FOR_TIMEZONE') ERROR at line 1:

ORA-01722: invalid number

原因：

ORACLE 11.2.0.3

SQL> select * from v$timezone_file;

FILENAME VERSION

RAC RMAN Restore to singe node

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

timezlrg_14.dat 14

ORACLE 11.2.0.1

SQL> select * from v$timezone_file;

FILENAME VERSION

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

timezlrg_11.dat 11

ORACLE 11.1.0.7

SQL> select * from v$timezone_file;

FILENAME VERSION

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

timezlrg.dat 4

ORACLE 10.2.0.4

SQL> select * from v$timezone_file;

FILENAME VERSION

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

timezlrg.dat 4

3. 执行utlu112s.sql脚本

此脚本显示升级过程中的一个摘要，不需要在upgrade模式下. SQL>@?/rdbms/admin/utlu112s.sql

4. 执行catuppst.sql脚本，同样不需要在upgrade模式下执行

此脚本用来迁移一些BASELINE数据到11g数据中.

SQL> @?/rdbms/admin/catuppst.sql

5. 编译无效对象

SQL> select count(*) from dba_invalid_objects;

SQL> @?/rdbms/admin/utlrp.sql

6. 检查组件状态

SQL> col comp_name for a40

SQL> set wrap off

本文来源：https://www.bwwdw.com/article/zeum.html