Oracle Dataguard之Real-Time Apply

Oracle Dataguard一共支持三种模式：最大可用模式（Maximum Availability），最大性能模式（Maximum Performance），最大保护模式（Maximum Protection）。默认创建的是最大性能模式（Maximum Performance）。关于三者的区别，我们来看官方的定义。

Maximum Availability

Transactions do not commit until all redo data needed to recover those transactions has been written to the online redo log and to the standby redo log on at least one synchronized standby database.If the primary database cannot write its redo stream to at least one synchronized standby database,it operates as if it were in maximum performance mode to preserve primary database availability until it is again able to write its redo stream to a synchronized standby database.

Maximum Performance

This is accomplished by allowing transactions to commit as soon as all redo data generated by those transactions has been written to the online log. Redo data is also written to one of more standby databases,but this is done asynchronously with respect to transaction commitment,so primary database performance is unaffected by delays in writing redo data to the standby databases.

Maximum Protection

This protection mode ensures that no data loss will occur if the primary database fails.To provide this level of protection,the redo data needed to recover a transaction must be written to both the online redo log and to the standby redo on at least one synchronized standby database before the transaction commits.To ensure that data loss cannot occur,the primary database will shut down,rather than continue processing transactions,if it cannot write its redo stream to at least one synchronized standby database.

三种模式对于日志传输的要求如下:

在前篇文章中-《Oracle Dataguard之物理standby的基本配置》，我们最后配置出来的是Physical Standby Database 最大性能模式下的异步传输，现在我们配置该模式下的Real-Time Apply。启用Real-Time Apply需要配置standby redo log，而这也是其它两种模式的基础。

一、创建standby redo log

1> 查询主库上online redo log的组数和大小 -->> node1 上操作

SQL> select group#,bytes/1024/1024 "size" from v$log;

GROUP# size
------ ----------
1 50
2 50
3 50

2> 在备库上添加standby redo log -->> group比主库多一组，大小和主库一样,node2上操作

[oracle@node2 ~]$ mkdir /u01/standby

SQL> alter database add standby logfile '/u01/standby/standby01.log' size 50M;

SQL> alter database add standby logfile '/u01/standby/standby02.log' size 50M;

SQL> alter database add standby logfile '/u01/standby/standby03.log' size 50M;

SQL> alter database add standby logfile '/u01/standby/standby04.log' size 50M;

注意：请确保已关闭Redo Apply,不然会报以下错误

SQL> alter database add standby logfile '/u01/standby/standby01.log' size 50M;

alter database add standby logfile '/u01/standby/standby01.log' size 50M
*
ERROR at line 1:
ORA-01156: recovery or flashback in progress may need access to files

关闭Redo Apply，

SQL> alter database recover managed standby database cancel;

二、修改主库的配置参数 -->> node1 上操作

SQL> alter system set log_archive_dest_2='service=to_victor lgwr affirm sync valid_for=(online_logfiles,primary_role) db_unique_name=victor';

SQL> alter system switch logfile;

三、在备库上启用Real-Time Apply -->> node2 上操作

SQL> alter database recover managed standby database using current logfile disconnect from session;

四、测试

除了用前篇文章中，查看归档日志的应用情况来验证dataguard的配置以外，本文将采用实际案例进行测试。

1> 在主库上新建一个测试表

SQL> create table test(id number);

SQL> insert into test values(1);

2> 在备库中检测

SQL> select * from test;

no rows selected

3> 在主库中提交事务

SQL> commit;

4> 在备库中检测

SQL> select * from test;

ID
----------
1

总结：

这个总结包括物理standby的基本配置和本文中的Real-Time Apply

1> 如果只设置主/次两个本地归档目的地，可以使用初始化参数log_archive_dest指定主归档目的地，使用初始化参数log_archive_duplex_dest指定次归档目的地。

在数据库中，初始换参数log_archive_dest和log_archive_duplex_dest与log_archive_dest_n只能使用一组来设置归档目的地，不能同时使用

2> 备库log_archive_dest_1如果没有显性指定，默认的归档目录将是$ORACLE_HOME/dbs.倘若显性指定，但valid_for不是

（standby_logfiles,standby_role）或者（all_logfiles,all_roles），则该设置无效，报警日志中将报以下错误：

ORA-16032: parameter STANDBY_ARCHIVE_DEST destination string cannot be translated

归档目录将继续为$ORACLE_HOME/dbs

3> 在本文中，log_archive_dest_1='location=/u01/archivelog valid_for=(standby_logfiles,standby_role) db_unique_name=victor'，如果再显性指定

log_archive_dest_3='location=/u01/standbyarchive valid_for=(standby_logfiles,standby_role) db_unique_name=victor'，将会有两份standby的

归档日志产生，没有必要，只需要一个log_archive_dest_1即可

4> 最大性能模式下，如果是async，即异步，则需要主库切一次日志，备库采用应用。而如果启用Real-Time Apply，则只需要主库事务commit，备库就能应用

5> 删除日志

SQL> alter database drop logfile group 7;

6> 如果发出了alter database recover managed standby database cancel;这个命令，MRP(Media Recovery process)将停止工作，但

RFS(Remote file server)仍继续工作

7> 如果没有standby redo logs，是不能启动real time apply的

SQL> alter database recover managed standby database using current logfile disconnect from session;
alter database recover managed standby database using current logfile disconnect from session
*
ERROR at line 1:
ORA-38500: USING CURRENT LOGFILE option not available without standby redo logs

8> standby数据库startup后，没有发出alter database recover managed standby database disconnect from session这个命令，RFS仍然工作，只要监听

都ok

9> disconnect from session代表这个命令后台运行。

10> Real-Time Apply的原理图