今天在hp的dp上，做了一次数据库的备份和恢复测试，大概的思路是这样的：
    1.备份策略:
        数据库全备份 backup database
        数据库归档日志备份 backup archivelog

    2.恢复策略
        a) 基于表空间的恢复
             shutdown immeidate;            
             mv data.dbf data.dbf.bak
             startup mount
             alter database open -> can't open db,because data.dbf not exist.
             use hp dp to restore tablespace data. 
             alter database open -> testing successfully.
        b) 数据库完全恢复
             shutdown immeidate;
             dd if=/dev/rlvdata66 of=/dev/rlvdata01 bs=8192k
             startup mount
             alter database open->can't open db, because rlvdata01 not exist.
             use hp dp to restore database , recover database.
             alter database open -> testing successfully.
         c)数据库不完全恢复(基于时间点的恢复)
            shutdown immediate;
            drop table test
            startup mount
            use hp dp to restore database to xxxx(time)
            alter database open resetlogs. 
            check select count(*) from test -> testing successfully.