关注公众号
备份:
Backup 有增量,copy 没有增量。
0,所有的使用的数据块,是基石
1—4 增量级别,备份<=n 以来的变化
1c---4c 累积增量,备份<=n-1 以来的变化
通过增量的级别来指定完备的备份策略
一、report:列出需要备份的所有物理文件--应该做什么
report schema;--数据库的结构是什么
report need backup days 3;
二、list:列出备份集详细信息--已经有什么了
list backup;
list backup of datafile 9;
list copy of controlfile;
list backup of controlfile;
list backup of spfile;
三、crosscheck:按照控制文件记录的信息,校验实际备份文件
crosscheck backup;
crosscheck copy;
delete –把控制文件中的信息删除
delete noprompt expired copy;
expired: rman的备份信息存储在控制文件,如果手工的删除备份文件,数据库不知道
如:
crosscheck archivelog all;
delete noprompt expired archivelog all ;
DELETE noprompt ARCHIVELOG ALL COMPLETED BEFORE 'SYSDATE-1';
四、show:查看参数
show all;
五、obsolete—不符合保留策略的;
report obsolete;
delete noprompt obsolete;
六、catalog命令
登记备份片:
catalog backuppiece '/setup/rman/f4.dbf';
登记目录下所有的备份片:
catalog start with '/setup/rman';
七、
增量备份:
backup incremental level 0 datafile 4 format ‘’;
累积增量备份:
backup cumulative incremental level 1 datafile 4 format ‘’;
八、控制文件备份:
(1)单独备份控制文件
(2)备份1号文件的时候,自动备份控制文件
(3)可以在备份的时候加上include current controlfile
backup datafile 4 include current controlfile format ‘/setup/bk/f4’;
(4)控制文件的自动备份策略
(5)控制文件的copy
backup current controlfile format ‘/setup/bk/%u.ctl’;
copy current controlfile to 'd:kc1.ctl';
copy current controlfile to '/setup/rman/1_bak.ctl';
相当于:
alter database backup controlfile to ‘’;
九、数据文件的备份:
backup datafile 4 format 'c:kf4_%u';
十、归档日志的备份:
copy archivelog 'F:oracleoradatazl9archARC264.LOG' to 'd:ka.cp';
BACKUP ARCHIVELOG ALL DELETE INPUT format 'd:karc%s.bk';
backup archivelog sequence between 264 and 265 thread=1 format 'd:karc%s.bk';
backup archivelog sequence between 50 and 52 thread=1 like '%0586360856%' format 'c:karc%s.bk';
十一、备份spfile:
backup spfile format 'd:kspfile.%s';
其他:
backup as compressed backupset database include current controlfile plus archivelog delete input format '/setup/rman/%T_%U';
backup database format 'c:k\%d_%s_%p' filesperset 3;
backup as compressed backupset format 'c:arcf4_%u' datafile 4 tag='users_full';
恢复:
一、数据块恢复
BLOCKRECOVER DATAFILE 12 BLOCK 12;
二、控制文件恢复:
1. 启动数据库到nomount状态
2.
rman target /
set dbid=XXX
如果controlfile全部丢失,可能无法获得dbid,需要使用restore controlfile from '';恢复控制文件。
RESTORE CONTROLFILE FROM 'C:kORA10O1_MF_NCNNF_TAG20071221T112712_3PPDNLWG_.BKP';
3. restore controlfile from autobackup;
4. alter database mount;
5. recover database;
6. alter database open resetlogs;
重新全备数据库。
三、数据文件的恢复:
select name,enabled,status from v$datafile;
1. alter database datafile 4 offline;
2. list backup of datafile 4 summary;
3. restore datafile 4;
4. recover datafile 4;
5. alter database datafile 4 online;
select name,enabled,status from v$datafile;
四、表空间的恢复:
restore tablespace users;
recover tablespace users;
alter tablespace users online;
select name,enabled,status from v$datafile;
五、系统表空间的恢复(undo,system,sysaux):
数据库启动到mount状态
restore tablespace system;
recover tablespace system;
alter database open;
六、不完全恢复
数据库得处于mount状态
使用scn:
run {
allocate channel c1 type DISK;
allocate channel c2 type DISK;
set until SCN = ****** ;
restore database;
recover database;
alter database open resetlogs; }
或
(scn号####)
recover database until change #### using backup controlfile;
使用时间点:
run {
allocate channel c1 type disk;
allocate channel c2 type disk;
sql 'alter session set nls_date_format="yyyy-mm-dd hh24:mi:ss"';
set until time = '2011-10-19 22:21:38';
restore database;
recover database;
alter database open resetlogs;
}
****************************
物理时间和数据库间的SCN 的对照表,每五分钟采样
select to_char(TIME_DP,'yyyy/mm/dd:hh24:mi:ss'),SCN from SYS.SMON_SCN_TIME;
set num 50
select current_scn from v$database;
****************************
直接恢复到最近:
recover database until cancel;
alter database open resetlogs;
七、数据库的完全恢复
启动到mount状态
Restore database;
Recover database;