ORACLE数据库的恢复(6)

SQL> conn internal 

SQL> shutdown immediate; 

 

用热备的控制文件恢复（把热备的控制文件拷贝回来） 

mount数据库 

SQL> startup mount 

ORACLE instance started. 

 

完全恢复和until cancel using backup controlfile都失败 

SQL> recover database; 

ORA-00283: recovery session canceled due to errors 

ORA-01610: recovery using the BACKUP CONTROLFILE option must be done 

 

SQL> recover database until cancel using backup controlfile; 

ORA-00279: change 424123 generated at 10/20/2002 20:44:12 needed for thread 1 

 

Specify log: {=suggested | filename | AUTO | CANCEL} 

auto 

ORA-00266: name of archived log file needed 

 

ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get error below 

ORA-01152: file 1 was not restored from a sufficiently old backup 

ORA-01110: data file 1: 'D:\BACKUPDB\SYSTEM01.DBF' 

 

重建控制文件  

SQL> alter database backup controlfile to trace; 

Database altered.

找到那个控制文件，然后编辑 

STARTUP NOMOUNT 

CREATE CONTROLFILE REUSE DATABASE "BACKUP" NORESETLOGS ARCHIVELOG 

    MAXLOGFILES 32 

    MAXLOGMEMBERS 2 

    MAXDATAFILES 254 

    MAXINSTANCES 1 

    MAXLOGHISTORY 453 

LOGFILE 

 GROUP 1 'D:\BACKUPDB\REDO01.LOG' SIZE 1M, 

 GROUP 2 'D:\BACKUPDB\REDO02.LOG' SIZE 1M, 

 GROUP 3 'D:\BACKUPDB\REDO03.LOG' SIZE 1M 

DATAFILE 

 'D:\BACKUPDB\SYSTEM01.DBF', 

 'D:\BACKUPDB\RBS01.DBF', 

 'D:\BACKUPDB\USERS01.DBF', 

 'D:\BACKUPDB\TEMP01.DBF', 

 'D:\BACKUPDB\TOOLS01.DBF', 

 'D:\BACKUPDB\INDX01.DBF' 

CHARACTER SET ZHS16GBK

; 

RECOVER DATABASE 

ALTER SYSTEM ARCHIVE LOG ALL; 

ALTER DATABASE OPEN; 

 

重建控制文件，并且恢复数据库（完全恢复成功！） 

SQL> shutdown immediate; 

ORA-01109: database not open

 

Database dismounted. 

ORACLE instance shut down. 

SQL> startup nomount; 

ORACLE instance started. 

 

SQL>@D:\BACKUPDB\udump\ORA02120.sql 

ORA-01081: cannot start already-running ORACLE - shut it down first 

Cluster altered. 

ORA-00283: recovery session canceled due to errors 

ORA-00264: no recovery required 

System altered. 

Database altered. 

 

验证恢复结果：完全恢复 

SQL> conn lunar/lunar 

SQL> select * from test1; 

完全恢复成功！ 

 

说明： 

当shutdown immediate的以后，如果丢失全部控制文件（不包括数据文件和redo），需要用热备的控制文件恢复数据库的时候，要想完全恢复（一直恢复到redo中commit的数据），必须执行以下步骤： 

1. mount数据库， 

2. backup controlfile to trace 

3. 修改这个生成的控制文件 

4. nomount 

5. 重建控制文件