Oracle - RMAN备份 之 incarnation的实验和小结
对于RMAN恢复我一直都不是很明白,因为,老是搞不清楚不完全恢复该怎么进行,今天,通过这个实验即是对不完全恢复的实践,
也是希望搞清楚incarnation到底是怎么工作的.很可惜,本人对Oracle的了解还不入门,仅能做一点提示.在文末做为小结.
RMAN> list backup;
List of Backup Sets
===================
BS Key Type LV Size Device Type Elapsed Time Completion Time
------- ---- -- ---------- ----------- ------------ -------------------
1 Full 559.92M DISK 00:01:08 2016-08-13 20:17:08 =============[第一个时间点]==========
BP Key: 1 Status: AVAILABLE Compressed: NO Tag: TAG20160813T201600
Piece Name: /home/oracle10/flash_recovery_area/DBINFO/backupset/2016_08_13/o1_mf_nnndf_TAG20160813T201600_cty3w0sx_.bkp
List of Datafiles in backup set 1
File LV Type Ckp SCN Ckp Time Name
---- -- ---- ---------- ------------------- ----
1 Full 1080165 2016-08-13 20:16:00 /home/oracle10/oradata/dbinfo/system01.dbf
2 Full 1080165 2016-08-13 20:16:00 /home/oracle10/oradata/dbinfo/undotbs01.dbf
3 Full 1080165 2016-08-13 20:16:00 /home/oracle10/oradata/dbinfo/sysaux01.dbf
4 Full 1080165 2016-08-13 20:16:00 /home/oracle10/oradata/dbinfo/users01.dbf
5 Full 1080165 2016-08-13 20:16:00 /home/oracle10/oradata/dbinfo/gias01.dbf
BS Key Type LV Size Device Type Elapsed Time Completion Time
------- ---- -- ---------- ----------- ------------ -------------------
2 Full 6.80M DISK 00:00:02 2016-08-13 20:17:17
BP Key: 2 Status: AVAILABLE Compressed: NO Tag: TAG20160813T201600
Piece Name: /home/oracle10/flash_recovery_area/DBINFO/backupset/2016_08_13/o1_mf_ncsnf_TAG20160813T201600_cty3yffw_.bkp
Control File Included: Ckp SCN: 1080190 Ckp time: 2016-08-13 20:17:15
SPFILE Included: Modification time: 2016-08-13 20:09:59
BS Key Type LV Size Device Type Elapsed Time Completion Time
------- ---- -- ---------- ----------- ------------ -------------------
3 Full 560.05M DISK 00:00:18 2016-08-13 20:26:33 =============[第二个时间点]==========
BP Key: 3 Status: AVAILABLE Compressed: NO Tag: TAG20160813T202615
Piece Name: /home/oracle10/orabak/fulldb_20160813_919801575
List of Datafiles in backup set 3
File LV Type Ckp SCN Ckp Time Name
---- -- ---- ---------- ------------------- ----
1 Full 1080933 2016-08-13 20:26:15 /home/oracle10/oradata/dbinfo/system01.dbf
2 Full 1080933 2016-08-13 20:26:15 /home/oracle10/oradata/dbinfo/undotbs01.dbf
3 Full 1080933 2016-08-13 20:26:15 /home/oracle10/oradata/dbinfo/sysaux01.dbf
4 Full 1080933 2016-08-13 20:26:15 /home/oracle10/oradata/dbinfo/users01.dbf
5 Full 1080933 2016-08-13 20:26:15 /home/oracle10/oradata/dbinfo/gias01.dbf
BS Key Type LV Size Device Type Elapsed Time Completion Time
------- ---- -- ---------- ----------- ------------ -------------------
4 Full 6.80M DISK 00:00:01 2016-08-13 20:26:41
BP Key: 4 Status: AVAILABLE Compressed: NO Tag: TAG20160813T202615
Piece Name: /home/oracle10/orabak/fulldb_20160813_919801600
Control File Included: Ckp SCN: 1080941 Ckp time: 2016-08-13 20:26:40
SPFILE Included: Modification time: 2016-08-13 20:09:59
BS Key Type LV Size Device Type Elapsed Time Completion Time
------- ---- -- ---------- ----------- ------------ -------------------
5 Full 561.88M DISK 00:00:19 2016-08-13 20:34:29 =============[第三个时间点]==========
BP Key: 5 Status: AVAILABLE Compressed: NO Tag: TAG20160813T203410
Piece Name: /home/oracle10/orabak/db_full_bak_20160813_919802050
List of Datafiles in backup set 5
File LV Type Ckp SCN Ckp Time Name
---- -- ---- ---------- ------------------- ----
1 Full 1081829 2016-08-13 20:34:10 /home/oracle10/oradata/dbinfo/system01.dbf
2 Full 1081829 2016-08-13 20:34:10 /home/oracle10/oradata/dbinfo/undotbs01.dbf
3 Full 1081829 2016-08-13 20:34:10 /home/oracle10/oradata/dbinfo/sysaux01.dbf
4 Full 1081829 2016-08-13 20:34:10 /home/oracle10/oradata/dbinfo/users01.dbf
5 Full 1081829 2016-08-13 20:34:10 /home/oracle10/oradata/dbinfo/gias01.dbf
BS Key Type LV Size Device Type Elapsed Time Completion Time
------- ---- -- ---------- ----------- ------------ -------------------
6 Full 6.80M DISK 00:00:01 2016-08-13 20:34:36
BP Key: 6 Status: AVAILABLE Compressed: NO Tag: TAG20160813T203410
Piece Name: /home/oracle10/orabak/db_full_bak_20160813_919802075
Control File Included: Ckp SCN: 1081837 Ckp time: 2016-08-13 20:34:35
SPFILE Included: Modification time: 2016-08-13 20:09:59
从上面查看备份集的输出可知,当前有三个可用的备份集.分别是:
1. 2016-08-13 20:17:08
2. 2016-08-13 20:26:33
3. 2016-08-13 20:34:29
现在的情况是,已经进行了一次完全恢复,使用的是 第3个时间点的全备.但现在需要恢复第2个时间点的全备.
注: 进行基于时间的备份时需要先导出指定日期格式的变量: export NLS_DATE_FORMAT='yyyy-mm-dd hh24:mi:ss';
RMAN> run {
set until time='2016-08-13 20:26:33';
restore database;
recover database;
alter database open resetlogs;
}
executing command: SET until clause
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of set command at 08/13/2016 23:42:18
RMAN-20207: UNTIL TIME or RECOVERY WINDOW is before RESETLOGS time
恢复时,出现了RMAN-20207的错误,这是进行恢复时,经常会出的错误.这是因为,我们先前恢复了第3个时间点
的全备,并进行了resetlogs,那么Oracle将的incarnation将加1,可以理解为Oracle将resetlogs前的备份集做为上层,
resetlogs后,重新建了一个层,开始重新记录信息. 我们现在要恢复上一层的备份集,就必须先回到上一层.
RMAN> list incarnation; #查看当前有几个incarnation.
List of Database Incarnations
DB Key Inc Key DB Name DB ID STATUS Reset SCN Reset Time
------- ------- -------- ---------------- --- ---------- ----------
1 1 DBINFO 3196490228 PARENT 1 2005-10-22 21:44:08
2 2 DBINFO 3196490228 PARENT 525876 2015-06-20 21:23:03
3 3 DBINFO 3196490228 CURRENT 1081970 2016-08-13 23:39:33 #这是当前层.
RMAN> reset database to incarnation 2;
RMAN> list incarnation;
List of Database Incarnations
DB Key Inc Key DB Name DB ID STATUS Reset SCN Reset Time
------- ------- -------- ---------------- --- ---------- ----------
1 1 DBINFO 3196490228 PARENT 1 2005-10-22 21:44:08
2 2 DBINFO 3196490228 CURRENT 525876 2015-06-20 21:23:03 #此时,它变成了当前层.
3 3 DBINFO 3196490228 ORPHAN 1081970 2016-08-13 23:39:33
4 4 DBINFO 3196490228 ORPHAN 1081972 2016-08-14 00:05:20
RMAN> run {
set until time='2016-08-13 20:26:33'; #使用第二次备份的时间点来进行不完全恢复.
restore database;
recover database;
alter database open resetlogs;
}
executing command: SET until clause
Starting restore at 2016-08-14 00:14:50
using channel ORA_DISK_1
.......
piece handle=/home/oracle10/orabak/fulldb_20160813_919801575 tag=TAG20160813T202615 #可以看到这里使用第二次的备份集.
channel ORA_DISK_1: restore complete, elapsed time: 00:00:55
Finished restore at 2016-08-14 00:15:45
恢复完成后,再次查看incarnation：
RMAN> list incarnation;
List of Database Incarnations
DB Key Inc Key DB Name DB ID STATUS Reset SCN Reset Time
------- ------- -------- ---------------- --- ---------- ----------
1 1 DBINFO 3196490228 PARENT 1 2005-10-22 21:44:08
2 2 DBINFO 3196490228 PARENT 525876 2015-06-20 21:23:03
5 5 DBINFO 3196490228 CURRENT 1080958 2016-08-14 00:15:47 #当前层在这里.
3 3 DBINFO 3196490228 ORPHAN 1081970 2016-08-13 23:39:33
4 4 DBINFO 3196490228 ORPHAN 1081972 2016-08-14 00:05:20
这时,我们继续将恢复的数据文件gias01.dbf 删除.然后,直接执行:
restore database; #发现恢复的是第二次备份的全备集.
recover database;
alter database open resetlogs; #并且,执行它时会报错,因为,Oracle不认为这是不完全恢复.
alter database open; #因此,直接此命令即可打开数据库.查看incarnation没有发生变化。
那现在要重新恢复第三次备份的备份集,要怎么做?
1. 重新关库,删除gias01.dbf
2. 重设incarnation 为4.
RMAN> reset database to incarnation 4;
RMAN> list incarnation;
List of Database Incarnations
DB Key Inc Key DB Name DB ID STATUS Reset SCN Reset Time
------- ------- -------- ---------------- --- ---------- ----------
1 1 DBINFO 3196490228 PARENT 1 2005-10-22 21:44:08
2 2 DBINFO 3196490228 PARENT 525876 2015-06-20 21:23:03
5 5 DBINFO 3196490228 ORPHAN 1080958 2016-08-14 00:15:47
3 3 DBINFO 3196490228 ORPHAN 1081970 2016-08-13 23:39:33
4 4 DBINFO 3196490228 CURRENT 1081972 2016-08-14 00:05:20 #可以看到现在已经修改成功.
3. 尝试手动恢复:
restore database; #这次恢复的是第三次备份的全备集.
recover database;
执行recover database时,出现下面错误，可忽略:
RMAN-03002: failure of recover command at 06/24/2015 16:02:25
RMAN-06054: media recovery requesting unknown log: thread 1 scn 277200603
4.进入sqlplus尝试打开数据库:
alter database open resetlogs; #成功打开.
5.再次查看incarnation：
RMAN> list incarnation;
List of Database Incarnations
DB Key Inc Key DB Name DB ID STATUS Reset SCN Reset Time
------- ------- -------- ---------------- --- ---------- ----------
1 1 DBINFO 3196490228 PARENT 1 2005-10-22 21:44:08
2 2 DBINFO 3196490228 PARENT 525876 2015-06-20 21:23:03
5 5 DBINFO 3196490228 ORPHAN 1080958 2016-08-14 00:15:47
3 3 DBINFO 3196490228 ORPHAN 1081970 2016-08-13 23:39:33
6 6 DBINFO 3196490228 CURRENT 1081972 2016-08-14 00:39:49 #现在第6层
4 4 DBINFO 3196490228 PARENT 1081972 2016-08-14 00:05:20
实验到此,我又尝试了reset database to incarnation 2; 然后,使用时间恢复第一次备份集,
成功,incarnation+1后为7,然后,我再次回到2,尝试使用时间恢复第一次备份,竟然失败了,然后,
我尝试使用时间恢复第一/二/三都不成功,但可以直接使用restore database来恢复.
我依然没有完全理解,从实验上看,没次resetlogs后,该层上将记录此次恢复备份集的信息,
并且上个当前层上关于该备份集的信息将被删除。
后记:
此次实验到最后,出现了一个问题,incarnation 被搞混乱了,我把所有的备份集都删除了,然后,
重新做了两次RMAN全备,每次都看incarnation有什么变化没,结果因为先前reset database to incarnation 3;
两次全备的信息似乎都没有记录到incarnation 3这层,无论我怎么切换,直接用restore database 或 基于时间的
恢复都提示找不到数据文件,将incarnation功能关闭也无法恢复.最终的解决方案是,重建控制文件来解决.
经测试重建控制文件可以重置incarnation.
这里有需要注意重建控制文件时: resetlogs与noresetlogs的区别
norestlogs: 控制文件的scn是来自当前日志的最大scn。
resetlogs: 控制文件的scn是来自数据文件。
重建控制文件的脚本:
1. 从$ORACLE_BASE/admin/$ORACLE_SID/bdump/alert_$ORACLE_SID.log ,从警告日志中找.
2. 手动写.
参考样本:
Create controlfile reuse database "dbinfo" noresetlogs archivelog
MAXINSTANCES 8
MAXLOGHISTORY 1
MAXLOGFILES 16
MAXLOGMEMBERS 3
MAXDATAFILES 100
LOGFILE
GROUP 1 '/home/oracle10/oradata/dbinfo/redo01.log' SIZE 51200K,
GROUP 2 '/home/oracle10/oradata/dbinfo/redo02.log' SIZE 51200K,
GROUP 3 '/home/oracle10/oradata/dbinfo/redo03.log' SIZE 51200K
-- STANDBY LOGFILE
Datafile
'/home/oracle10/oradata/dbinfo/system01.dbf',
'/home/oracle10/oradata/dbinfo/undotbs01.dbf',
'/home/oracle10/oradata/dbinfo/sysaux01.dbf',
'/home/oracle10/oradata/dbinfo/users01.dbf',
'/home/oracle10/oradata/dbinfo/gias01.dbf'
CHARACTER SET ZHS16GBK
;
写好重建控制文件的脚本后,启动Oracle到nomount：
SQL> startup nomount
SQL> @/tmp/create-control-file.sql
