oracle rman 异机恢复pitr的具体步骤

rman 使用备份集进行异机还原、恢复。大家应该清楚还原和恢复的差异。

$ rman target  /

/*
测试用，发现测试失败，如果需要还原spfile时就很有用
没有参数文件，在RMAN的命令下会生成一个DUMMY的实例
RMAN> set dbid=1234567890;
*/
启动数据库到nomunt状态

RMAN> startup nomount pfile='/ebsfs/EBSPROD2/db/tech_st/11.2.0/dbs/initEBSPROD.ora';

还原控制文件

RMAN> run {
  ALLOCATE CHANNEL ch00 TYPE disk;
  restore controlfile to '/ebsfs/EBSPROD2/db/apps_st/data/cntrl01.dbf' from '/ebstar/ebsclonetar2/CTL_c-xxxxxxxx-04';
  restore controlfile to '/ebsfs/EBSPROD2/db/apps_st/data/cntrl02.dbf' from '/ebstar/ebsclonetar2/CTL_c-xxxxxxxx-04';
  restore controlfile to '/ebsfs/EBSPROD2/db/apps_st/data/cntrl03.dbf' from '/ebstar/ebsclonetar2/CTL_c-xxxxxxxx-04';
  RELEASE CHANNEL ch00;
}

修改数据库到mount状态

RMAN> alter database mount;

效验备份集

RMAN> crosscheck backupset;

删除备份集

RMAN> delete expired backup;

注册本地目录,一定要最后一个’/’

RMAN>catalog start with '/ebstar/ebsclonetar2/';

做还原的检查

RMAN> run{
   ALLOCATE CHANNEL ch00 TYPE disk;
   restore database preview summary;
   RELEASE CHANNEL ch00;
}

Media recovery start SCN is xxxxxxxx
Recovery must be done beyond SCN xxxxxxxx to clear datafile fuzziness
Finished restore at xx-xxx-xx

released channel: ch00

新开一个shell会话,拼接还原SQL

$ Sqlplus /nolog 

SQL> conn / as sysdba;
Connected.

SQL> set lines 500;
SQL> set pages 500;

SQL> select 'set newname for datafile '''||dd.name||''' to '''||replace(replace(dd.name,'/EBSPROD/','/EBSPROD2/'),'/ebsfs2/','/ebsfs/')||''';' as set_dbname
from v$datafile dd
where 1=1
order by dd.name
; 

还原数据库,一般开4个channel是比较快了。如果io不错，可以开6个、8个channel试试。
把上面输出的语句粘贴到下面 ———- 处

RMAN> run{
  ALLOCATE CHANNEL ch00 TYPE disk;
  ALLOCATE CHANNEL ch01 TYPE disk;
  ALLOCATE CHANNEL ch02 TYPE disk;
  ALLOCATE CHANNEL ch03 TYPE disk;
  ----------
  ----------
  restore database;
  switch datafile all;
  RELEASE CHANNEL ch00;
  RELEASE CHANNEL ch01;
  RELEASE CHANNEL ch02;
  RELEASE CHANNEL ch03;
}

尽可能多的应用归档，切记不要贸然打开数据库

RMAN> run{
  ALLOCATE CHANNEL ch00 TYPE disk;
  recover database ;
  RELEASE CHANNEL ch00;
}

或者指定一个时间点
RMAN> run{
  ALLOCATE CHANNEL ch00 TYPE disk;
  set until time "to_date('2006-12-01 15:00:00','yyyy-mm-dd hh24:mi:ss')";
  recover database ;
  RELEASE CHANNEL ch00;
}

结束后会报一个ora错误提示无法找到指定的归档文件
这时可以从源库拷贝生成的归档文件

再次注册本地目录,一定要最后一个’/’

RMAN> catalog start with '/ebstar/ebsclonetar2/';
RMAN> list backup summary;
RMAN> list backup ;

再次恢复数据库

RMAN> run{
  ALLOCATE CHANNEL ch00 TYPE disk;
  recover database ;
  RELEASE CHANNEL ch00;
}

新开一个sqlplus会话，修改文件

sqlplus /nolog

SQL*Plus: Release 11.2.0.1.0 Production on Mon Feb 24 09:34:38 2016

Copyright (c) 1982, 2009, Oracle.  All rights reserved.

SQL> conn / as sysdba;
Connected.

SQL> select instance_name,status from v$instance;


SQL> set lines 500;
SQL> set pages 500;

SQL> select 'alter database rename file '''||dd.file_name||''' to '''||replace(replace(dd.file_name,'/EBSPROD/','/EBSPROD2/'),'/ebsfs2/','/ebsfs/')  ||''';' as alter_tmpname
from (
      select name as file_name from v$tempfile
      union all
      select member as file_name from v$logfile l 
      ) dd
Order by dd.file_name

拷出来执行就行了

打开数据库

SQL> select instance_name,status from v$instance;

SQL> alter database open resetlogs;

调整TEMP 表空间

select 'alter database tempfile '''||dtf.file_name||''' autoextend on next 100m maxsize 30g; ', dtf.*
from dba_temp_files dtf
where 1=1
;

关闭数据库的归档模式

如果源数据库是RAC，目标恢复机是单实例的话还需要做如下操作

禁止thread

SQL> select thread#,status,enabled from v$thread;
SQL> alter database disable thread 2;

删除被禁止thread 的 logfile group

SQL> select group#,thread#,archived,status from v$log;
SQL> alter database clear unarchived logfile group 5;
SQL> alter database drop logfile group 5;

清除多余的undo

SQL> select name from v$tablespace where name like 'UNDO%';
SQL> show parameter undo_tablespace;
NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
undo_tablespace                      string      UNDOTBS1

SQL> drop tablespace undotbs2 including contents and datafiles;

Tablespace dropped.

再检查tnsnames.ora 及 listener.ora