近日接手一套新500GB的  9i  dataguard架构的DBServer、HP380

由于IO瓶颈、导致在做rman backup时用户反映产线‘卡死’，选择在standbydb上备份测量。

       

        方案实施步骤：

         一 、备份standbydb

              RMAN> run{

                                      delete obsolete;

                                      allocate channel db01 device type disk format '/data2/rmanbackup/database_%T_%U.bak';

                                      allocate channel db02 device type disk format '/data2/rmanbackup/database_%T_%U.bak';

                                      allocate channel db03 device type disk format '/data2/rmanbackup/database_%T_%U.bak';

                                      allocate channel db04 device type disk format '/data2/rmanbackup/database_%T_%U.bak';

                                      BACKUP DATABASE  filesperset 5;

                                      release channel db01;

                                    }

       二、copy standby controlfile

               # cp /u01/app/oracle/oradata/ygzhou/control01.ctl   /data2/rmanbackup/control01.ctl

      三、利用standby backupset/standby controlfile  restore database

               RMAN> restore controlfile from '/ data2/rmanbackup/control01.ctl';

               RMAN> run{

                                 restore database;

                                    }

     四、switchower or failover this standbydb to primarydb

 

注意：1、standby备份、primarydb的控制文件不存在该备份信息，所以利用primarydb 的控制文件用常规的方式无法恢复database;

            2、由于在rman backup过程中DB中会有大量的dml或者ddl (drop语句可能导致备份失败) 出现，导致controlfile中的scn号与datafile中的scn不匹配；此后需求利用到archivelog来保证数据安全  （archivelog的重要性....）。