月归档：五月 2011

前两天公司的一套oracle dg因为维护人员的疏忽到值出现gap，无法恢复，只能重做，因为该库每天都有rman的增量备份，所以选择最近的一次全部进行重做dg

1、主库
ALTER DATABASE CREATE STANDBY CONTROLFILE AS ‘/tmp/control01.ctl’;
alter system archive log current;

2、修改pfile文件（因为我是重做，所以不需要做）
见11g data guard（PHYSICAL STANDBY）配置

3、拷贝控制文件/rman文件到备库的相同位置
备份全备的内容（/opt/ORBS/data /tmp/control01.ctl)

4、复制数据库做dg
启动到nomount状态
rman target sys/srtsysdb@zjch_Action auxiliary /
duplicate target database for standby dorecover nofilenamecheck;

出现以下错误
RMAN-06025: no backup of log thread 1 seq 675 lowscn 1096519 found to restore
RMAN-06025: no backup of log thread 1 seq 674 lowscn 1095666 found to restore
RMAN-06025: no backup of log thread 1 seq 673 lowscn 1095663 found to restore
RMAN-06025: no backup of log thread 1 seq 672 lowscn 1095656 found to restore
RMAN-06025: no backup of log thread 1 seq 671 lowscn 1095081 found to restore
RMAN-06025: no backup of log thread 1 seq 670 lowscn 1092545 found to restore
RMAN-06025: no backup of log thread 1 seq 669 lowscn 1090351 found to restore
RMAN-06025: no backup of log thread 1 seq 668 lowscn 1088837 found to restore
RMAN-06025: no backup of log thread 1 seq 667 lowscn 1088834 found to restore
RMAN-06025: no backup of log thread 1 seq 666 lowscn 1088721 found to restore
原因：因为控制文件和备份文件中有时间间隔，所以rman要恢复，但是归档日志没有传输过来，所以不能恢复。

–到主库的归档目录查看这些日志是否存在
ls -l（如果存在，下面的操作恢复日志的操作可以忽略）
–查看这些日志是否存在
list backup of archivelog from sequence 666 until sequence 675;
–如果存在rman备份中，使用rman进行恢复
restore archivelog from sequence 666 until sequence 675;

如果存在相关联机日志就不需要添加
–添加redo log
alter database add logfile group 1 (‘/opt/oracle/oradata/zjch/redo01.log’) size 50m;
alter database add logfile group 2 (‘/opt/oracle/oradata/zjch/redo02.log’) size 50m;
alter database add logfile group 3 (‘/opt/oracle/oradata/zjch/redo03.log’) size 50m;
–添加sandby log
ALTER DATABASE ADD STANDBY LOGFILE (‘/opt/oracle/oradata/zjch/std_redo04.log’) SIZE 50M reuse;
ALTER DATABASE ADD STANDBY LOGFILE (‘/opt/oracle/oradata/zjch/std_redo05.log’) SIZE 50M reuse;
ALTER DATABASE ADD STANDBY LOGFILE (‘/opt/oracle/oradata/zjch/std_redo06.log’) SIZE 50M reuse;
ALTER DATABASE ADD STANDBY LOGFILE (‘/opt/oracle/oradata/zjch/std_redo07.log’) SIZE 50M reuse;

5、开启日志应用
–开启实时应用
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE USING CURRENT LOGFILE DISCONNECT FROM SESSION;
–开启应用（根据配置和需求）
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE DISCONNECT FROM SESSION;
到这一步会如果主机端有所需的归档日志，会自动传输过来，并进行恢复，不用人工干预。

–查看alter文件，查询日志传输是否正常
–查看日志是否传输完成
select thread#, low_sequence#, high_sequence# from v$archive_gap;

–主库切换日志
alter system switch logfile;
–查询最大sequence
select max(sequence#) from v$archived_log;

–备库查看alert文件
tail -100 alert.log
Media Recovery Log /opt/oracle/oradata/zjch/archive/stdarch/1_46544_681145105.dbf
Media Recovery Log /opt/oracle/oradata/zjch/archive/stdarch/1_46545_681145105.dbf 有类似内容
–查询最大sequence
select max(sequence#) from v$archived_log;

–主库建立一个测试表
create table SCOTT.t_test(id number,name varchar2(10));
insert into SCOTT.t_test values(1,’aaaa’);
commit;
alter system switch logfile;
–查询最大sequence
select max(sequence#) from v$archived_log;

–备库检查
–查询最大sequence
select max(sequence#) from v$archived_log;
–取消日志应用
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE cancel;
–只读模式打开数据库
alter database open read only;
–查询刚刚建的表和插入数据是否通报不成功
select * from SCOTT.t_test;

–重启备库，设置为日志应用模式
shutdown immediate;
startup mount;
–开启实时应用
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE USING CURRENT LOGFILE DISCONNECT FROM SESSION;
–开启应用（根据配置和需求）
ALTER DATABASE RECOVER MANAGED STANDBY DATABASE DISCONNECT FROM SESSION;

–主库删除测试表
drop table SCOTT.t_test;