Oracle
資料庫做異地恢復時，全庫的完全恢復需要很大的空間和很長的時間。
如果採用基於模式的恢復方式僅僅恢復某一模式下所有物件，而不用去恢復所有的資料，這將節省大量的時間和空間。
這個在
Oracle 9i
版本中經常使用，在
10g
、
11g
中也可以使用一下。
下面案例是在
10g
的環境下將一個
ASM
上的
RAC
資料庫的資料恢復到一個檔案系統上單實例資料庫中，並且是只恢復幾個使用者下的資料。
操作步驟如下：
1
、將備份組拷貝到要恢復的機器上，如果空間夠的話，可以將資料備份、控制檔案備份和歸檔備份一起拷貝過來，如果不夠的話，先拷貝控制檔案備份和資料檔案備份。
2
、關閉資料庫，啟動到
nomount
狀態進行控制檔的恢復
shutdown immediate;
startup nomount;
rman target /
restore controlfile from '/data/urpdb/urpdb_ctl_ikn3tot4_1_1.20120222';
3
、修改控制檔
查詢要恢復的使用者的表空間資訊
select owner,tablespace_name,count(*) from dba_segments group by owner,tablespace_name;
查找這些表空間和系統表空間（
'SYSTEM','SYSAUX', 'UNDOTBS1','UNDOTBS2', 'UNDOTBS3', 'UNDOTBS4'
）對應的資料檔案
select *
from dba_data_files
where tablespace_name in
('TSP_URP', 'TSP_URP_INDEX', 'TSP_JC', 'TSP_JC_INDEX', 'TSP_GXSJ',
'TSP_GXSJ_INDE', 'TSP_OWB', 'TSP_YJSSJCK', 'USERS', 'SYSTEM',
'SYSAUX', 'UNDOTBS1','UNDOTBS2', 'UNDOTBS3', 'UNDOTBS4');
SQL> alter database backup controlfile to trace;
修改控制檔並執行：
CREATE CONTROLFILE REUSE DATABASE "URPDB" RESETLOGS FORCE LOGGING ARCHIVELOG
MAXLOGFILES 192
MAXLOGMEMBERS 3
MAXDATAFILES 1024
MAXINSTANCES 32
MAXLOGHISTORY 9088
LOGFILE
GROUP 1 (
'/opt/app/oracle/oradata/URPDB/group1_2.log',
'/opt/app/oracle/oradata/URPDB/group1_1.log'
) SIZE 10M,
GROUP 2 (
'/opt/app/oracle/oradata/URPDB/group2_2.log',
'/opt/app/oracle/oradata/URPDB/group2_1.log'
) SIZE 10M,
GROUP 3 (
'/opt/app/oracle/oradata/URPDB/group3_2.log',
'/opt/app/oracle/oradata/URPDB/group3_1.log'
) SIZE 10M,
GROUP 11 '/opt/app/oracle/oradata/URPDB/group11_1.log' SIZE 100M,
GROUP 12 '/opt/app/oracle/oradata/URPDB/group12_1.log' SIZE 100M,
GROUP 13 '/opt/app/oracle/oradata/URPDB/group13_1.log' SIZE 100M,
GROUP 14 '/opt/app/oracle/oradata/URPDB/group14_1.log' SIZE 100M,
GROUP 15 '/opt/app/oracle/oradata/URPDB/group15_1.log' SIZE 100M
-- STANDBY LOGFILE
DATAFILE
'/opt/app/oracle/oradata/URPDB/system01.dbf',
'/opt/app/oracle/oradata/URPDB/undotbs1.dbf',
'/opt/app/oracle/oradata/URPDB/sysaux.dbf',
'/opt/app/oracle/oradata/URPDB/users.dbf',
'/opt/app/oracle/oradata/URPDB/undotbs2.dbf',
'/opt/app/oracle/oradata/URPDB/tsp_urp.dbf',
'/opt/app/oracle/oradata/URPDB/tsp_urp_index.dbf',
'/opt/app/oracle/oradata/URPDB/tsp_gxsj.dbf',
'/opt/app/oracle/oradata/URPDB/tsp_urp.dbf_1.ora',
'/opt/app/oracle/oradata/URPDB/tsp_owb.ora',
'/opt/app/oracle/oradata/URPDB/tsp_jc.dbf',
'/opt/app/oracle/oradata/URPDB/tsp_jc_index.dbf',
'/opt/app/oracle/oradata/URPDB/tsp_yjssjck.dbf',
'/opt/app/oracle/oradata/URPDB/undotbs03.dbf',
'/opt/app/oracle/oradata/URPDB/undotbs04.dbf'
CHARACTER SET AL32UTF8
;
mount
資料庫；
SQL> alter database mount;
select name from v$datafile;
--select 'alter database rename file '''||name||''' to '''|| replace(name,'+URPDBDG','/opt/app/oracle/oradata/URPDB')||''';' from v$datafile;
3
、將備份資料檔案拷貝到測試機
rman>catalog start with '/data/urpdb/';
執行一下，就是將這個目錄下的資料檔案的備份檔案寫入了
catalog
。
4
、
restore
資料庫
select file#,name from v$datafile
；
---
看看控制檔中資料檔案存放的位置，這裡是
“
+URPDBDG/..."
，替換為目標資料庫檔存放的位置。
set linesize 300
column name format a200
set pagesize 100
select 'set newname for datafile '||file#||' to '''||replace(name,'+URPDBDG','/opt/app/oracle/oradata/urpdb')||''';' newname from v$datafile --
如果一個目標目錄不夠，可以用多個目標目錄
union all
select 'restore datafile '||file#||';' newname from v$datafile;
select *
from dba_data_files
where tablespace_name in
('TSP_URP', 'TSP_URP_INDEX', 'TSP_JC', 'TSP_JC_INDEX', 'TSP_GXSJ',
'TSP_GXSJ_INDE', 'TSP_OWB', 'TSP_YJSSJCK', 'USERS', 'SYSTEM',
'SYSAUX', 'UNDOTBS1','UNDOTBS2', 'UNDOTBS3', 'UNDOTBS4');
rman>
run
{
allocate channel ch01 type disk;
allocate channel ch02 type disk;
set until time "to_date('2012-02-23 10:00:00','yyyy-mm-dd hh24:mi:ss')";
set newname for datafile 1 to '/opt/app/oracle/oradata/URPDB/system01.dbf';
set newname for datafile 2 to '/opt/app/oracle/oradata/URPDB/undotbs1.dbf';
set newname for datafile 3 to '/opt/app/oracle/oradata/URPDB/sysaux.dbf';
set newname for datafile 4 to '/opt/app/oracle/oradata/URPDB/users.dbf';
set newname for datafile 6 to '/opt/app/oracle/oradata/URPDB/undotbs2.dbf';
set newname for datafile 9 to '/opt/app/oracle/oradata/URPDB/tsp_urp.dbf';
set newname for datafile 10 to '/opt/app/oracle/oradata/URPDB/tsp_urp_index.dbf';
set newname for datafile 14 to '/opt/app/oracle/oradata/URPDB/tsp_gxsj.dbf';
set newname for datafile 16 to '/opt/app/oracle/oradata/URPDB/tsp_urp.dbf_1.ora';
set newname for datafile 18 to '/opt/app/oracle/oradata/URPDB/tsp_owb.ora';
set newname for datafile 19 to '/opt/app/oracle/oradata/URPDB/tsp_jc.dbf';
set newname for datafile 20 to '/opt/app/oracle/oradata/URPDB/tsp_jc_index.dbf';
set newname for datafile 26 to '/opt/app/oracle/oradata/URPDB/tsp_yjssjck.dbf';
set newname for datafile 29 to '/opt/app/oracle/oradata/URPDB/undotbs03.dbf';
set newname for datafile 30 to '/opt/app/oracle/oradata/URPDB/undotbs04.dbf';
restore datafile 1;
restore datafile 2;
restore datafile 3;
restore datafile 4;
restore datafile 6;
restore datafile 9;
restore datafile 10;
restore datafile 14;
restore datafile 16;
restore datafile 18;
restore datafile 19;
restore datafile 20;
restore datafile 26;
restore datafile 29;
restore datafile 30;
switch datafile all;
release channel ch01;
release channel ch02;
}
5
、修改日誌檔和暫存檔案的路徑為目標資料庫的路徑（可與
restore
同時進行）
select member from v$logfile;
select name from v$tempfile;
select 'alter database rename file '''||member||''' to '''|| replace(member,'+URPDBDG','/opt/app/oracle/oradata/URPDB')||''';' from v$logfile;
alter database rename file '+YWKDG/ywk/onlinelog/group15_1.log' to '/opt/app/oracle/oradata2/YWKDG/group15_1.log';
。。。。。
select 'alter database rename file '''||name||''' to '''|| replace(name,'+URPDBDG','/opt/app/oracle/oradata2/urpdb/')||''';' from v$tempfile;
alter database rename file '+YWKDG/ywk/tempfile/tsp_zc_temp02.dbf' to '/opt/app/oracle/oradata2/YWKDG/tsp_zc_temp02.dbf';
6
、將歸檔檔拷到目標資料庫伺服器，
rman>catalog start with '/data/urpdb/';
執行一下，就是將這個目錄下的歸檔檔的備份檔案寫入了
catalog
。
run
{
allocate channel ch01 type disk;
allocate channel ch02 type disk;
set until time "to_date('2012-02-23 10:00:00','yyyy-mm-dd hh24:mi:ss')";
recover database;
release channel ch01;
release channel ch02;
}
或則
SQL> select to_char(checkpoint_time,'yyyy-mm-dd hh24:mi:ss') from v$datafile;
S
QL> select file#,checkpoint_change# from v$datafile order by checkpoint_change#;
FILE# CHECKPOINT_CHANGE#
---------- ------------------
14 2793109724
20 2793109725
6 2793232002
19 2793232002
18 2793258889
26 2793258889
9 2793281171
30 2793281171
1 2793283158
16 2793301522
29 2793301522
2 2793304369
10 2793304369
4 2793333238
3 2793333238
15 rows selected.
run
{
allocate channel ch01 type disk;
allocate channel ch02 type disk;
set until scn 2793333238;
recover database
；
release channel ch01;
release channel ch02;
}
如果
recover
資料庫缺少檔，可以到主庫去查是哪個檔，然後拷貝過來再
recover
一次。
RMAN> list backupset of archivelog logseq 7438 thread 4;
RMAN> crosscheck backupset;
delete expired backupset;
7
、
alter database open resetlogs;
扫一扫
1238
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
