ORACLE 12C 冷备份方式的数据迁移（重建控制文件）

实验对象：两台linux单机的oracle 12C 数据库（大版本和小版本都一致），数据迁移采取冷备份的方式。目标机器的路径跟源库不一致，需要重建controlfile。

迁移步骤

首先关闭监听，和kill掉连接的应用
ps -ef | grep LOCAL=NO | awk '{print ($2)}' | xargs kill -9

创建pfile
create pfile='/tmp/inittest.ora1012' from spfile;

查看数据文件，日志文件的路径，日志文件路径

SQL> set line 200
SQL> col FILE_NAME for a80

SQL> select TABLESPACE_NAME,file_name from dba_data_files;

TABLESPACE_NAME FILE_NAME
------------------------------ --------------------------------------------------------------------------------
SYSTEM /u01/app/oracle/oradata/test/system01.dbf
SYSAUX /u01/app/oracle/oradata/test/sysaux01.dbf
UNDOTBS1 /u01/app/oracle/oradata/test/undotbs01.dbf
USERS /u01/app/oracle/oradata/test/users01.dbf
QWERTY /u01/app/oracle/oradata/test/qwerty.dbf
TESTBIG /u01/app/oracle/oradata/test/testbig.dbf
DATA /u01/app/oracle/oradata/test/data_01.dbf

SQL> select TABLESPACE_NAME,file_name from dba_temp_files;

TABLESPACE_NAME FILE_NAME
------------------------------ --------------------------------------------------------------------------------
TEMP /u01/app/oracle/oradata/test/temp01.dbf
TEMP_ASYNC /u01/app/oracle/oradata/test/temp_async_01.dbf

SQL> set line 200
SQL> col MEMBER for a80
SQL> select GROUP#,MEMBER from v$logfile;

GROUP# MEMBER
---------- --------------------------------------------------------------------------------
1 /u01/app/oracle/oradata/test/redo01.log
2 /u01/app/oracle/oradata/test/redo02.log
3 /u01/app/oracle/oradata/test/redo03.log

SQL> show parameter control_files

NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
control_files string /u01/app/oracle/oradata/test/c
ontrol01.ctl, /u01/app/oracle/
fast_recovery_area/test/contro
l02.ctl

关闭数据库后，拷贝参数文件，数据文件，日志文件，控制文件。
shutdown immediate

将所有需要的文件拷到本地，以便更好的传输。
cp ......

拷贝文件到目标服务器上。
scp ...

在目标服务器上修改pfile的信息。

源库
[oracle@test dbs]$ vi inittest.ora
test.__data_transfer_cache_size=0
test.__db_cache_size=339738624
test.__java_pool_size=4194304
test.__large_pool_size=8388608
test.__oracle_base='/u01/app/oracle'#ORACLE_BASE set from environment
test.__pga_aggregate_target=293601280
test.__sga_target=545259520
test.__shared_io_pool_size=16777216
test.__shared_pool_size=167772160
test.__streams_pool_size=0
*.audit_file_dest='/u01/app/oracle/admin/test/adump'
*.audit_trail='db'
*.compatible='12.1.0.2.0'
*.control_files='/u01/app/oracle/oradata/test/control01.ctl'
*.db_block_size=8192
*.db_domain=''
*.db_name='test'
*.db_recovery_file_dest='/u01/app/oracle/fast_recovery_area'
*.db_recovery_file_dest_size=4815m
*.diagnostic_dest='/u01/app/oracle'
*.dispatchers='(PROTOCOL=TCP) (SERVICE=testXDB)'
*.log_archive_format='%t_%s_%r.dbf'
*.memory_target=800m
*.open_cursors=300
*.processes=300
*.remote_login_passwordfile='EXCLUSIVE'
*.undo_tablespace='UNDOTBS1'

以下操作在目标库上执行:

修改了控制文件的路径
*.control_files='/u01/app/oracle/oradata/test/controlfile/control01.ctl'

重建控制文件
startup mount

alter database backup controlfile to trace as '/tmp/crontol_trace';

cat /tmp/crontol_trace | grep -v ^- | grep -v ^$ > /tmp/ctl.sql

修改控制文件里的路径,因为我们有onlinelog所以选择noresetlog模式
vi /tmp/ctl.sql

STARTUP NOMOUNT
CREATE CONTROLFILE REUSE DATABASE "TEST" NORESETLOGS ARCHIVELOG
MAXLOGFILES 16
MAXLOGMEMBERS 3
MAXDATAFILES 100
MAXINSTANCES 8
MAXLOGHISTORY 292
LOGFILE
GROUP 1 '/u01/app/oracle/oradata/test/onlinelog/redo01.log' SIZE 50M BLOCKSIZE 512,
GROUP 2 '/u01/app/oracle/oradata/test/onlinelog/redo02.log' SIZE 50M BLOCKSIZE 512,
GROUP 3 '/u01/app/oracle/oradata/test/onlinelog/redo03.log' SIZE 50M BLOCKSIZE 512
DATAFILE
'/u01/app/oracle/oradata/test/datafile/system01.dbf',
'/u01/app/oracle/oradata/test/datafile/sysaux01.dbf',
'/u01/app/oracle/oradata/test/datafile/undotbs01.dbf',
'/u01/app/oracle/oradata/test/datafile/users01.dbf',
'/u01/app/oracle/oradata/test/datafile/qwerty.dbf',
'/u01/app/oracle/oradata/test/datafile/testbig.dbf',
'/u01/app/oracle/oradata/test/datafile/data_01.dbf'
CHARACTER SET AL32UTF8
;
RECOVER DATABASE
ALTER SYSTEM ARCHIVE LOG ALL;
ALTER DATABASE OPEN;
ALTER TABLESPACE TEMP ADD TEMPFILE '/u01/app/oracle/oradata/test/datafile/temp01.dbf' REUSE;
ALTER TABLESPACE TEMP ADD TEMPFILE '/u01/app/oracle/oradata/test/datafile/temp02.dbf' REUSE;
ALTER TABLESPACE TEMP_ASYNC ADD TEMPFILE '/u01/app/oracle/oradata/test/datafile/temp_async_01.dbf' REUSE;

修改完毕后，将数据库启动到nomount状态
shutdown immediate

startup nomout;

运行重建控制文件的sql语句
@/tmp/ctl.sql

SQL> select status from v$instance;

STATUS
------------
OPEN

至此，已经完成数据库的迁移。

本文出自 “Sysdba” 博客，请务必保留此出处http://sysdba.blog.51cto.com/10492366/1702149