DataGuard开启延时应用的测试

实验环境：RHEL 6.5 + Oracle 11.2.0.4 GI、DB + Primary RAC（2 nodes）+ Standby RAC（2 nodes）

1.体验DG延时应用的效果

主库 log_archive_dest_N 参数当前设定值：

log_archive_dest_2                   string      SERVICE=mynas VALID_FOR=(ONLIN
                                                 E_LOGFILES,PRIMARY_ROLE) DB_UN
                                                 IQUE_NAME=mynas

我需要将其修改，增加 DELAY=N 分钟，我这里为了试验方便，设置了5分钟，实际一般都需要根据业务要求延时1天（1440）甚至更久。

alter system set log_archive_dest_2 = 'SERVICE=mynas DELAY=5 VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=mynas';

设置完成之后，如果备库之前是USING CURRENT LOGFILE实时应用，那么还需要将备库应用停止重新以不加USING CURRENT LOGFILE方式启动应用，否则将会忽略这个DELAY的参数，这一点我们从主库和备库的日志均可以看到。

主库日志：

Tue Aug 15 15:19:05 2017
Archived Log entry 741 added for thread 1 sequence 287 ID 0x97764e10 dest 1:
Tue Aug 15 15:19:05 2017
WARNING: Managed Standby Recovery started with USING CURRENT LOGFILE
  DELAY 5 minutes specified at primary ignored
ARC0: Standby redo logfile selected for thread 1 sequence 287 for destination LOG_ARCHIVE_DEST_2

备库日志：

Tue Aug 15 15:18:28 2017
Media Recovery Waiting for thread 1 sequence 287
Tue Aug 15 15:19:05 2017
WARNING: Managed Standby Recovery started with REAL TIME APPLY
  DELAY 5 minutes specified at primary ignored
RFS[33]: Assigned to RFS process 5482
RFS[33]: Selected log 11 for thread 1 sequence 287 dbid -1785877518 branch 919999037
Tue Aug 15 15:19:06 2017
Archived Log entry 353 added for thread 1 sequence 287 ID 0x97764e10 dest 1:
ARC3: Archive log thread 1 sequence 287 available in 5 minute(s)
Tue Aug 15 15:19:06 2017
Managed Standby Recovery started with USING CURRENT LOGFILE
Ignoring previously specified DELAY 5 minutes for thread 1 sequence 287
Media Recovery Log +FRA/mynas/archivelog/2017_08_15/thread_1_seq_287.563.952096745
Media Recovery Waiting for thread 2 sequence 235

注：我建立一个每秒都会插入一条数据的测试表，可用来更清楚的去辅助验证延迟应用的设置是否生效。

因此, 需要将备库取消应用，再以不加USING CURRENT LOGFILE的方式重新开启应用，命令如下：

alter database recover managed standby database cancel;
alter database recover managed standby database disconnect from session;

再观察主库日志：

Tue Aug 15 15:33:22 2017
Archived Log entry 743 added for thread 1 sequence 288 ID 0x97764e10 dest 1:
Tue Aug 15 15:33:22 2017
ARC0: Standby redo logfile selected for thread 1 sequence 288 for destination LOG_ARCHIVE_DEST_2

备库日志：

Tue Aug 15 15:33:22 2017
RFS[33]: Selected log 11 for thread 1 sequence 288 dbid -1785877518 branch 919999037
Tue Aug 15 15:33:23 2017
Archived Log entry 354 added for thread 1 sequence 288 ID 0x97764e10 dest 1:
ARC2: Archive log thread 1 sequence 288 available in 4 minute(s)

可以看到延迟应用的设置已经生效。

Tue Aug 15 15:36:11 2017
RFS[33]: Selected log 11 for thread 1 sequence 289 dbid -1785877518 branch 919999037
Tue Aug 15 15:36:12 2017
Archived Log entry 355 added for thread 1 sequence 289 ID 0x97764e10 dest 1:
ARC3: Archive log thread 1 sequence 289 available in 5 minute(s)
Tue Aug 15 15:36:14 2017
RFS[34]: Assigned to RFS process 6921
RFS[34]: Selected log 21 for thread 2 sequence 235 dbid -1785877518 branch 919999037
Tue Aug 15 15:36:15 2017
Archived Log entry 356 added for thread 2 sequence 235 ID 0x97764e10 dest 1:
ARC0: Archive log thread 2 sequence 235 available in 3 minute(s)
Tue Aug 15 15:36:20 2017
Media Recovery Delayed for 3 minute(s) (thread 2 sequence 235)
Tue Aug 15 15:39:17 2017
Media Recovery Log +FRA/mynas/archivelog/2017_08_15/thread_2_seq_235.566.952097775
Media Recovery Log +FRA/mynas/archivelog/2017_08_15/thread_1_seq_287.563.952096745
Media Recovery Log +FRA/mynas/archivelog/2017_08_15/thread_1_seq_288.564.952097603
Media Recovery Delayed for 5 minute(s) (thread 1 sequence 289)

注：我这里同时观察测试表的数据，可以看到，实际上延迟应用，是保证达到设定的值之后才能应用对应的归档日志，本身还会有延迟。当备库日志显示刚刚应用日志时，这时的差距是最小的，基本和设置的时间一致。

2.总结：开启、关闭延时应用

开启延时应用方法：主库直接动态修改参数log_archive_dest_N，加入DELAY=（单位是分钟）的参数，然后备库确保不加USING CURRENT LOGFILE启动应用。

--主库：
alter system set log_archive_dest_2 = 'SERVICE=mynas DELAY=5 VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=mynas';


--备库：
alter database recover managed standby database cancel;
alter database recover managed standby database disconnect from session;

那么如何从延时应用重新改回实时应用呢？就是再设置回去就ok了。

--主库：
alter system set log_archive_dest_2 = 'SERVICE=mynas VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=mynas';

--备库：
alter database recover managed standby database cancel;
alter database recover managed standby database using current logfile disconnect from session;

注：由于整个操作涉及的参数是动态可以修改的，所以整个过程都可以在线完成，不需要停机。

3.测试表脚本

每隔1s插入一条数据的测试表

--业务用户jingyu下创建序列和测试表test
conn jingyu/jingyu
create sequence jingyu.s1;
drop table jingyu.test;
create table jingyu.TEST(id number, insert_time timestamp);

--需要sys用户授权dbms_lock给业务用户jingyu
grant execute on dbms_lock to jingyu;

--创建每隔1s插入一条数据的存储过程，持续10小时
create or replace procedure p1 is
begin
for i in  1..36000
loop
insert into jingyu.test select s1.nextval, systimestamp from dual;
commit;
dbms_lock.sleep(1);
end loop;
end;
/

--编辑执行存储过程的脚本
[oracle@jyrac1 ~]$ vi insert.sh
sqlplus jingyu/jingyu <<EOF
exec p1;
EOF

--后台执行脚本
[oracle@jyrac1 ~]$ nohup sh insert.sh &