使用using current logfile实现DG备库实时更新

时间:2021-04-19 17:45:00

DG1是主库,DG2是备库;运行在最大可用模式。

实验一:未使用using current logfile参数时,备库在最大可用模式下,不能实时同步备库的更新

-此时需要等待主库进行归档---侧面证明备库应用的归档来恢复数据。

BYS@dg1>show parameter log_archive_dest_2
NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_2                   string      SERVICE=dg2 LGWR SYNC AFFIRM V
                                                 ALID_FOR=(ONLINE_LOGFILES,PRIM
                                                 ARY_ROLE) DB_UNIQUE_NAME=dg2
主库上的log_archive_dest_2 参数中LGWR SYNC AFFIRM意思是用LGWR传输日志,SYNC是同步传输,ARRIFRM是需要确认。
日志的传输方法有以下 :
使用ARCH–把完整的REDO文件COPY到standby数据库服务器上。
使用LGWR,有两种模式:
ASYNC - 异步:redo written by LGWR to local disk---LGWR将redo写到本地LOG文件。LNSn进程读取redo,传到standby服务器上。
SYNC - 同步:Redo written to standby by LGWR -  LGWR讲Redo信息直接写到standby数据库服务器上
[oracle@dg2 ~]$ sqlplus /  as sysdba                 
BYS@dg2>select protection_mode,database_role,protection_level from v$database;
PROTECTION_MODE      DATABASE_ROLE    PROTECTION_LEVEL
-------------------- ---------------- --------------------
MAXIMUM AVAILABILITY PHYSICAL STANDBY MAXIMUM AVAILABILITY
SYS@dg2>alter database recover managed standby database disconnect from session;
Database altered.

1.此时在dg1-主库上进行插入数据操作

BYS@dg1>select * from test;
         A
----------
       999
       123
       333
       111
BYS@dg1>insert into test values(555);
1 row created.
BYS@dg1>commit;
Commit complete.

2.返回dg2备库上进行查询操作--验证数据同步情况

我这里未使用ACTIVE STANDBY,如果使用更方便,只需要停止应用并查询就可以了。

SYS@dg2>alter database recover managed standby database cancel;
Database altered.
SYS@dg2>alter database open;
Database altered.
SYS@dg2>select open_mode from v$database;
OPEN_MODE
--------------------
READ ONLY
SYS@dg2>conn bys/bys
Connected.
不能查询到DG1上的最新数据。
BYS@dg2>select * from test;
         A
----------
       999
       123
       333
       111
BYS@dg2>conn / as sysdba
SYS@dg2>shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.
SYS@dg2>startup mount;

3.此时切换主库REDO日志,产生的日志大致如下:--日志是后来补上的,日期不 一样,环境 相同

[oracle@dg1 ~]$ cat alert_dg.log
Thu Aug 08 10:07:43 2013
LGWR: Standby redo logfile selected to archive thread 1 sequence 44
LGWR: Standby redo logfile selected for thread 1 sequence 44 for destination LOG_ARCHIVE_DEST_2
Thread 1 advanced to log sequence 44 (LGWR switch)
  Current log# 2 seq# 44 mem# 0: /u01/oradata/dg/redo02.log
Thu Aug 08 10:07:43 2013
Archived Log entry 65 added for thread 1 sequence 43 ID 0x676c9833 dest 1:   -------43号归档已经完成

[oracle@dg2 ~]$ cat alert_dg.log
Thu Aug 08 10:07:42 2013
Standby controlfile consistent with primary
RFS[1]: Selected log 4 for thread 1 sequence 44 dbid 1735160627 branch 821829622
Thu Aug 08 10:07:42 2013
Archived Log entry 29 added for thread 1 sequence 43 ID 0x676c9833 dest 1:   -------43号归档已经完成
Thu Aug 08 10:07:46 2013
Media Recovery Log /u01/archivelog/arc_1_43_821829622.arc
Media Recovery Waiting for thread 1 sequence 44 (in transit)

实验二:使用using current logfile参数,可以实时同步主库的更新。

SYS@dg2>alter database
recover managed standby database using current logfile disconnect from session;
Database altered.

1.此时在主库上进行DML操作:

BYS@dg1>
BYS@dg1>set time on
17:31:36 BYS@dg1>insert into test values(666);
1 row created.
17:31:49 BYS@dg1>
17:31:54 BYS@dg1>commit;
Commit complete.
17:31:58 BYS@dg1>

2.在备库上以read-only模式打开数据库并进行查看--要在主库的操作提交后再到备库上操作:

SYS@dg2>set time on
17:31:39 SYS@dg2>
17:32:11 SYS@dg2>alter database recover managed standby database cancel;
Database altered.
17:32:40 SYS@dg2>alter database open;
Database altered.
17:32:54 SYS@dg2>conn bys/bys
Connected.
查询结果显示,备库上已经同步到了主库的最新操作。
17:32:59 BYS@dg2>select * from test;
         A
----------
       999
       123
       333
       555
       666
       111
17:33:03 BYS@dg2>

3.此时的日志大致如下,也有日志切换(alter system switch lofile)

11:15:39 SYS@dg1>alter system switch logfile;
System altered.
11:15:48 SYS@dg1>exit

[oracle@dg1 ~]$ cat alert_dg.log
Thu Aug 08 11:15:48 2013
LGWR: Standby redo logfile selected to archive thread 1 sequence 55
LGWR: Standby redo logfile selected for thread 1 sequence 55 for destination LOG_ARCHIVE_DEST_2
Thread 1 advanced to log sequence 55 (LGWR switch)
  Current log# 1 seq# 55 mem# 0: /u01/oradata/dg/redo01.log
Thu Aug 08 11:15:48 2013
Archived Log entry 87 added for thread 1 sequence 54 ID 0x676c9833 dest 1:   54号归档已经完成

[oracle@dg2 ~]$ cat alert_dg.log
Thu Aug 08 11:15:48 2013
Archived Log entry 40 added for thread 1 sequence 54 ID 0x676c9833 dest 1:   54号归档已经完成
Thu Aug 08 11:15:48 2013
Standby controlfile consistent with primary
Thu Aug 08 11:15:48 2013
Media Recovery Waiting for thread 1 sequence 55 (in transit)
RFS[12]: Selected log 4 for thread 1 sequence 55 dbid 1735160627 branch 821829622
Recovery of Online Redo Log: Thread 1 Group 4 Seq 55 Reading mem 0
  Mem# 0: /u01/oradata/dg/standbyredo04.log
Thu Aug 08 11:15:49 2013
RFS[13]: Assigned to RFS process 3881
RFS[13]: Identified database type as 'physical standby': Client is ARCH pid 2929

http://blog.csdn.net/haibusuanyun/article/details/11524053