不停机 data guard 及 switchover
DATA GUARD实施文档：
环境描述：
服务器1：一个运行的OLTP数据库，不能停机。数据库安装归属人 ora ，ORACLE_SID=research
服务器2：一个已经安装了数据库的服务器，有一个数据库在跑，不能影响。数据库安装归属人 oracle ，ORACLE_SID=orcl
服务器基本环境
| Hostmachine IP |
OS release |
sys passwd |
database release |
database SID |
databse unique_name |
DG role |
installers |
| 192.168.0.15 |
Red Hat 5 |
oracle |
oracle Enterprise 10.2.0.1.0 |
orcl |
orcl |
standby |
oracle |
| 192.168.0.16 |
Red Hat 5 |
oracle |
oracle Enterprise 10.2.0.1.0 |
research |
research |
primary |
ora |
实施步骤：
DG我们的主备都是用SPFILE启动实例的。
primary端：
一、强制数据库logging和开启archivelog
SQL> archive log list;
SQL> shutdown immediate;
SQL> startup mount ;
SQL> alter database archivelog;
SQL> archive log list;
SQL> alter database force logging;
SQL> select force_logging from v$database;
SQL> altet database open;
二、 创建或是修改OS密码文件：
orapwd file=$ORACLE_HOME/dbs/orapw$ORACLE_SID password=oracle entries=30 force=y
然后SCP到standby端去。 （注意要修改ORACLE_SID 因为我们这里的SID 不同。直接MV 就可以了。）
三、备份控制文件：
Sql>alter database create standby controlfie as ‘/u01/backup/control.ctl’;或是RMAN 备份
RMAN>backup current controlfile for standby format =‘/u01/backup/control.ctl’;（我是用RMAN全库备份带上controlfile 所以没有单独进行备份）。然后SCP给standby注意复制成多份，并MV到parameter file 指定位置。
四、修改listener.ora 和 tnsname.ora
Listener.ora 加上静态监听 （主备都加上静态监听）
# listener.ora Network Configuration File: /opt/oracle/product/10.2.0/db_1/network/admin/listener.ora
# Generated by Oracle configuration tools.
SID_LIST_LISTENER =
(SID_LIST =
(SID_DESC =
(SID_NAME = PLSExtProc)
(ORACLE_HOME = /opt/oracle/product/10.2.0/db_1)
(PROGRAM = extproc)
)
(SID_DESC =
(GLOBAL_DBNAME = research.bj)
(ORACLE_HOME = /opt/oracle/product/10.2.0/db_1)
(SID_NAME = research)
)
)
LISTENER =
(DESCRIPTION_LIST =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC1))
(ADDRESS = (PROTOCOL = TCP)(HOST = dg-pp)(PORT = 1521))
)
)
这里注意细节，SID_DESC 不能顶格，还有括号不能少哦！GLOBAL_DBNAME=service_name
Tnsname.ora
# tnsnames.ora Network Configuration File: /opt/oracle/product/10.2.0/db_1/network/admin/tnsnames.ora
# Generated by Oracle configuration tools.
EXTPROC_CONNECTION_DATA =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC1))
)
(CONNECT_DATA =
(SID = PLSExtProc)
(PRESENTATION = RO)
)
)
orcl =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.0.16)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = orcl)
)
)
research =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.0.15)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = research.bj)
)
)
这里注意orcl必须顶格，否则报错
五、修改parameter file
research.__db_cache_size=310378496
research.__java_pool_size=4194304
research.__large_pool_size=4194304
research.__shared_pool_size=134217728
research.__streams_pool_size=0
*.audit_file_dest='/opt/oracle/admin/research/adump'
*.background_dump_dest='/opt/oracle/admin/research/bdump'
*.compatible='10.2.0.1.0'
*.control_files='/opt/oracle/oradata/research/control01.ctl','/opt/oracle/oradata/research/control02.ctl','/opt/oracle/oradata/research/control03.ctl'
*.core_dump_dest='/opt/oracle/admin/research/cdump'
*.db_block_size=8192
*.db_domain=''
*.db_file_multiblock_read_count=8
*.db_name='research'
*.db_recovery_file_dest='/opt/oracle/flash_recovery_area'
*.db_recovery_file_dest_size=2147483648
*.dispatchers='(PROTOCOL=TCP) (SERVICE=researchXDB)'
*.job_queue_processes=10
*.open_cursors=300
*.pga_aggregate_target=150994944
*.processes=150
*.remote_login_passwordfile='EXCLUSIVE'
*.sga_target=455081984
*.undo_management='AUTO'
*.undo_tablespace='UNDOTBS1'
*.user_dump_dest='/opt/oracle/admin/research/udump'
####以下是主库修改的参数部分###
*.fal_client='research'
*.fal_server='orcl'
*.log_archive_config='DG_CONFIG=(research,orcl)'
*.log_archive_dest_1='LOCATION=/arch VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=research'
*.log_archive_dest_2='SERVICE=orcl LGWR ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=orcl'
*.log_archive_dest_state_1='ENABLE'
*.log_archive_dest_state_2='DEFER'
*.standby_file_management='AUTO'
因为我们是在线搭建DG，所以我们就直接命令修改了，但是DB_UNIQUE_NAME需要重启才能生效，所以我们就按默认research好了。
以下为主库在线修改命令：
alter system set fal_client='research';
alter system set fal_server='orcl';
alter system set log_archive_config='DG_CONFIG=(research,orcl)';
alter system set log_archive_dest_1='LOCATION=/arch VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=research';
alter system set log_archive_dest_2='SERVICE=orcl LGWR ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=orcl';
alter system set log_archive_dest_state_1='ENABLE';
alter system set log_archive_dest_state_2='DEFER'; ##（这里为什么要DEFER，因为ENABLE自动传输会照成我们RMAN全库备份长时间等待，ENABLE他就会传输，可是咱们还没有搭建成功呢，所以这里先DEFER，等我们DG搭建好，我们在修改成ENABLE，然日志输出）##
alter system set standby_file_management='AUTO';
六、RMAN备份全库
RUN {
allocate channel c1 type disk;
allocate channel c2 type disk;
sql 'alter system archive log current';
backup current controlfile for standby format='/u01/backup/control_%U';
BACKUP FORMAT '/u01/backup/orcl_%U_%T' skip inaccessible filesperset 5 DATABASE ;
sql 'alter system archive log current';
BACKUP FORMAT '/u01/backup/arch_%U_%T' skip inaccessible filesperset 5 ARCHIVELOG ALL DELETE INPUT;
release channel c2;
release channel c1;
}
过程省略，然后把备份文件SCP 到 standby相同的目录下，如果不同会出现ORA-19505
到这里primary端操作基本完成，下面我们来standby
*********************************************************************
*********************************************************************
standby端
一、 修改OS密码文件：
primary端的 OS密码文件scp过来，并修改成standby的 ORACLE_SID
二、修改parameter file：
primary端的parameter file 文件 scp过来，并修改成standby的 ORACLE_SID
Primary file 修改
oracle.__db_cache_size=310378496(这里注意我们scp过来的是research.修改成oracle.)
oracle.__java_pool_size=4194304
oracle.__large_pool_size=4194304
oracle.__shared_pool_size=134217728
oracle.__streams_pool_size=0
*.audit_file_dest='/opt/oracle/admin/research/adump'
*.background_dump_dest='/opt/oracle/admin/research/bdump'
*.compatible='10.2.0.1.0'
*.control_files='/opt/oracle/oradata/research/control01.ctl','/opt/oracle/oradata/research/control02.ctl','/opt/oracle/oradata/research/control03.ctl'
*.core_dump_dest='/opt/oracle/admin/research/cdump'
*.db_block_size=8192
*.db_domain=''
*.db_file_multiblock_read_count=8
*.db_name='research'
*.db_recovery_file_dest='/opt/oracle/flash_recovery_area'
*.db_recovery_file_dest_size=2147483648
*.dispatchers='(PROTOCOL=TCP) (SERVICE=researchXDB)'
*.job_queue_processes=10
*.log_archive_dest_1='location=/u01/arch'
*.open_cursors=300
*.pga_aggregate_target=150994944
*.processes=150
*.remote_login_passwordfile='EXCLUSIVE'
*.sga_target=455081984
*.undo_management='AUTO'
*.undo_tablespace='UNDOTBS1'
*.user_dump_dest='/opt/oracle/admin/research/udump'
#### add standby parameter file ###
*.DB_UNIQUE_NAME='orcl'
*.LOG_ARCHIVE_CONFIG='DG_CONFIG=(orcl,research)'
*.LOG_ARCHIVE_DEST_1='LOCATION=/u01/arch VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=orcl'
*.LOG_ARCHIVE_DEST_2='SERVICE=research LGWR ASYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=research'
*.LOG_ARCHIVE_DEST_STATE_1=ENABLE
*.LOG_ARCHIVE_DEST_STATE_1=ENABLE
*.FAL_SERVER='research'
*.FAL_CLIENT='orcl'
*.STANDBY_FILE_MANAGEMENT='AUTO'
如果我们准备的数据文件目录，和admin目录不同，那我们需要在参数文件内指出并修改(这里我们主库和备库的目录相同所以没有修改目录，也方便了我们RMAN恢复的时候，不用RENAME了)。
三、修改控制文件：
primary端的控制文件scp过来后，要修改成参数文件制定的名字和目录。否则参数文件找不到，要把一份复制成多份，具体几份那就看你的参数文件了。
（我们这里是用RMAN直接恢复的，我们的目录也没有改变所以就不用指定了）。
四、修改listener.ora和tnsname.ora
Listener.ora 加上静态监听
# listener.ora Network Configuration File: /opt/oracle/product/10g/network/admin/listener.ora
# Generated by Oracle configuration tools.
SID_LIST_LISTENER =
(SID_LIST =
(SID_DESC =
(SID_NAME = PLSExtProc)
(ORACLE_HOME = /opt/oracle/product/10g)
(PROGRAM = extproc)
)
(SID_DESC =
(GLOBAL_DBNAME = orcl)
(ORACLE_HOME = /opt/oracle/product/10g)
(SID_NAME = orcl)
)
)
LISTENER =
(DESCRIPTION_LIST =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC1))
(ADDRESS = (PROTOCOL = TCP)(HOST = dg-ss)(PORT = 1521))
)
)
Tnsname.ora 修改
# tnsnames.ora Network Configuration File: /opt/oracle/product/10g/network/admin/tnsnames.ora
# Generated by Oracle configuration tools.
RMAN =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = dg-ss)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = rman)
)
)
EXTPROC_CONNECTION_DATA =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC1))
)
(CONNECT_DATA =
(SID = PLSExtProc)
(PRESENTATION = RO)
)
)
orcl =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.0.16)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = orcl)
)
)
research =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.0.15)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = research.bj)
)
)
Lsnrctl stop
Lsnrctl start
Lsnrclt status
这里我们配置了2个，主要目的就是切换的时候不用在配置了。这里主备的相同的。如果备库没有network/admin的文件，可以从主库cp过来。TNSPING 互相必须能PING通。
五、recover standby database
1）把主库备份的备份文件，scp到相同的目录下，如果目录不同需要注册。
2）Standby 端需要启动到nomount 状态，且用spfile启动。（具体过程省略）。
3）利用RMAN 恢复standby database
Primary 端：
[ora@dg-pp dbs]$rman target / auxiliary sys/system@orcle
_s
恢复管理器: Release 10.2.0.4.0 - Production on 星期五 5月 27 15:33:05 2011
Copyright (c) 1982, 2007, Oracle. All rights reserved.
连接到目标数据库: DGDB1 (DBID=152223982)
已连接到辅助数据库: DGDB1 (未装载)
RMAN> duplicate target database for standby nofilenamecheck dorecover;
1. 这里需要解释一下我们直接用了 NOFILENAMECHECK 参数，是因为我们的主库，备库的数据文件目录相同。不同的话，需要我们制定脚本。
2. RUN{ # set new file names for the datafiles SET NEWNAME FOR DATAFILE 1 TO '?/dbs/standby_data_01.f'; SET NEWNAME FOR DATAFILE 2 TO '?/dbs/standby_data_02.f'; # run the DUPLICATE command DUPLICATE TARGET DATABASE FOR STANDBY;}
接下来就等待了，等到RMAN RECOVER 结束。查看STANDBY状态，为MOUNT。这里一般情况会提示 未完成恢复，因为我们做DG这些操作的时候必须耗费时间，这段时间内我们没有备份主库的归档，也就是说你做完全库后，主库在运行还会产生归档的。这里恢复的时候会提示，但是这不用担心，我们的目的不就是做 DG ，DG的意义就在于把没有同步过来的数据进行同步，我们的DG搭建好后。没有被RMAN恢复的归档自然会被DG传送过来了。
Standby 端：
SQL*Plus: Release 10.2.0.1.0 - Production on Sun Apr 1 17:58:41 2012
Copyright (c) 1982, 2005, Oracle. All rights reserved.
SQL> conn /as sysdba
Connected.
SQL> select status from v$instance;
STATUS
------------------------
MOUNTED
六、创建STANDYB REDO
因为我们使用的 LOG ASYNC ROLE 所以我们需要创建 standby redo
这里需要在主备都创建standby redo 且大小，组数，必须相同。
Standby redo = (onredo+1) * 线程数 线程数这里可以认为是RAC的节点数。
我们这里是单实例，所以创建了4组 standby redo。
SQL> ALTER DATABASE ADD STANDBY LOGFILE GROUP 4（’/数据文件目录/redo04.log）size 50M; (一共创建4组，步骤省略)
Primary 端：
在主库也创建哦！
SQL> ALTER DATABASE ADD STANDBY LOGFILE GROUP 4（’/数据文件目录/redo04.log）size 50M; (一共创建4组，步骤省略)
接下来有点混乱一会primary端 一会standby端 注意好哦
standby端：
确保主库OPEN，备库MOUNT。
1、从库将接收从主库传送过来的日志
primary端：
SQL> alter system set log_archive_dest_state_2='ENABLE';
System altered.
Standby 端：
2、接收并应用日志。
SQL> alter database recover managed standby database disconnect from session;
System altered.
SQL> select switchover_status from v$database;
SWITCHOVER_STATUS
----------------------------------------
SESSIONS ACTIVE
这里sessions active 有session活动，这种情况正常。即使出现 NOT ALLOWED 也不要害怕，因为这些状态不能确定你的DG 是否成功。因为SWITCHOVER_STATUS仅仅表示DG转换的状态。
到这里standby端的操作完成了。
*********************************************************************
*********************************************************************
日志传送和应用测试
测试日志有没有传送和同步
1、主库
SQL>select max(sequence#) from v$archived_log;
SQL>alter system switch logfile;
SQL>select max(sequence#) from v$archived_log;
2、从库
SQL>select max(sequence#) from v$archived_log;
两个库上看到的序列号一致表示日志成功传送；
在从库上应用日志
SQL>alter database recover managed standby database disconnect from session;
测试日志是否能应用（测试数据是否一致）
1、主库
在主库上建立一个测试表并插入数据并提交 比如：
SQL>conn scott/tigger
SQL>create table test (name varchar2(10),job varchar2(20));
SQL>insert into test values('qiujunhua','dba');
SQL>insert into test values('qjh','plsql developer');
SQL>commit;
SQL>alter system switch logfile;
2、从库
在从库上应用日志
SQL>alter database recover managed standby database disconnect from session;
在open从库之前先停止应用日志
SQL>alter database recover managed standby database cancel;
打开数据库验证数据是否一致
SQL>alter database open;
SQL>conn scott/tigger
SQL>select * from test;
若看到的数据和主库一致，说明日志成功应用
3、返回到接收日志状态
从库启动应用日志:（在从库OPEN的状态下直接执行不用先shutdown)
SQL>alter database recover managed standby database disconnect from session;
从库停止应用日志，但是继续接收日志
SQL>alter database recover managed standby database cancel;
执行后从库处于mount状态
*********************************************************************
*********************************************************************
Data guard switchover
查看主库，备库的状态。
primary端：
SQL> select switchover_status from v$database;
SWITCHOVER_STATUS
--------------------
SESSIONS ACTIVE （证明还有session活动）
standby端：
SQL> select switchover_status from v$database;
SWITCHOVER_STATUS
----------------------------------------
NOT ALLOWED （备库等待主库经行操作，此状态不影响switchover）
2、进行switchover 操作：
primary端：
SQL> select switchover_status from v$database;
SWITCHOVER_STATUS
--------------------
SESSIONS ACTIVE
如果此状态为’to standby’ 则可以直接切换
使用“alter database commit to switchover to physical standby ;”
我们当前状态为‘sessions active’ 说明还有session 活动。
使用“alter database commit to switchover to physical standbywith session shutdown;” 执行完这个命令，数据库为nomount状态。
SQL> alter database commit to switchover to physical standby with session shutdown;
Database altered.
SQL> select status from v$instance;
STATUS
------------
STARTED
执行这个命令后，我们来看看 alert 日志。这样可以让我们知道 命令都干了些什么
alter database commit to switchover to physical standby with session shutdown
Sun Apr 1 18:49:21 2012
ALTER DATABASE COMMIT TO SWITCHOVER TO PHYSICAL STANDBY (orcl)
Sun Apr 1 18:49:25 2012
Thread 1 advanced to log sequence 31
Current log# 2 seq# 31 mem# 0: /opt/oracle/oradata/research/redo02.log
Sun Apr 1 18:49:25 2012
ARCH: Standby redo logfile selected for thread 1 sequence 30 for destination LOG_ARCHIVE_DEST_2
Sun Apr 1 18:49:27 2012
Stopping background process CJQ0
Sun Apr 1 18:49:27 2012
SMON: disabling tx recovery
Sun Apr 1 18:49:27 2012
Stopping background process QMNC
Sun Apr 1 18:49:28 2012
Thread 1 advanced to log sequence 32
Current log# 3 seq# 32 mem# 0: /opt/oracle/oradata/research/redo03.log
Sun Apr 1 18:49:28 2012
ARC0: Standby redo logfile selected for thread 1 sequence 31 for destination LOG_ARCHIVE_DEST_2
Sun Apr 1 18:49:28 2012
Stopping Job queue slave processes
Sun Apr 1 18:49:36 2012
Waiting for Job queue slaves to complete
Sun Apr 1 18:50:03 2012
Job queue slave processes stopped
All dispatchers and shared servers shutdown
Sun Apr 1 18:50:05 2012
SMON: disabling cache recovery
Sun Apr 1 18:50:05 2012
Shutting down archive processes
Archiving is disabled
Sun Apr 1 18:50:10 2012
ARCH shutting down
ARC1: Archival stopped
Sun Apr 1 18:50:15 2012
ARC0: Becoming the heartbeat ARCH
ARC0: Archiving disabled
ARCH shutting down
ARC0: Archival stopped
Sun Apr 1 18:50:16 2012
Thread 1 closed at log sequence 32
Successful close of redo thread 1
Sun Apr 1 18:50:16 2012
ARCH: Noswitch archival of thread 1, sequence 32
ARCH: End-Of-Redo Branch archival of thread 1 sequence 32
ARCH: Archiving is disabled due to current logfile archival
Clearing standby activation ID 2102571270 (0x7d52b106)
The primary database controlfile was created using the
'MAXLOGFILES 16' clause.
There is space for up to 13 standby redo logfiles
Use the following SQL commands on the standby database to create
standby redo logfiles that match the primary database:
ALTER DATABASE ADD STANDBY LOGFILE 'srl1.f' SIZE 52428800;
ALTER DATABASE ADD STANDBY LOGFILE 'srl2.f' SIZE 52428800;
ALTER DATABASE ADD STANDBY LOGFILE 'srl3.f' SIZE 52428800;
ALTER DATABASE ADD STANDBY LOGFILE 'srl4.f' SIZE 52428800;
Archivelog for thread 1 sequence 32 required for standby recovery
MRP0 started with pid=10, OS id=17838
Sun Apr 1 18:50:17 2012
MRP0: Background Managed Standby Recovery process started (orcl)
Managed Standby Recovery not using Real Time Apply
Online logfile pre-clearing operation disabled by switchover
Media Recovery Log /u01/arch/1_32_779474103.dbf
Identified End-Of-Redo for thread 1 sequence 32
Sun Apr 1 18:50:22 2012
Media Recovery End-Of-Redo indicator encountered
Sun Apr 1 18:50:22 2012
Media Recovery Applied until change 500306
Sun Apr 1 18:50:22 2012
MRP0: Media Recovery Complete: End-Of-REDO (orcl)
Resetting standby activation ID 2102571270 (0x7d52b106)
Sun Apr 1 18:50:23 2012
Waiting for MRP0 pid 17838 to terminate
Sun Apr 1 18:50:24 2012
MRP0: Background Media Recovery process shutdown (orcl)
Sun Apr 1 18:50:24 2012
idle dispatcher 'D000' terminated, pid = (13, 1)
Sun Apr 1 18:50:24 2012
Switchover: Complete - Database shutdown required (orcl)
Sun Apr 1 18:50:24 2012
Completed: alter database commit to switchover to physical standby with session shutdown
此时看备库的状态已经成为 switchover pending 。这足以证明我们之前看到standby状态为not allowed 只是等待主库给她发状态命令罢了。
SQL> select switchover_status from v$database;
SWITCHOVER_STATUS
----------------------------------------
SWITCHOVER PENDING
此时主库已切换成了，备库。现在重启到mount
Shut immediate
Startup mount
SQL> select switchover_status from v$database;
SWITCHOVER_STATUS
--------------------
TO PRIMARY
此时的状态也表明了我们转换成功了，现在这个是备库，备库的switchover 状态为to primary 。
到这里primary 已经转换成 standby了。
Standby 端：
1、上面的 primary 已经转换成功了，我们看一下standby状态。
SQL> select switchover_status from v$database;
SWITCHOVER_STATUS
----------------------------------------
SWITCHOVER PENDING
这里需要我们应用日志了。
使用“alter database recover managed standby database disconnect from session;”
SQL> alter database recover managed standby database disconnect from session;
Database altered.
SQL> select switchover_status from v$database;
SWITCHOVER_STATUS
----------------------------------------
TO PRIMARY
此时standby的状态时to primary 允许我们转换了。
使用“alter database commit to switchover to primary;”如果执行此命令后hold住的话，你可以使用下面命令经行操作。
SQL> alter database recover managed standby database disconnect from session;
alter database recover managed standby database disconnect from session
*
ERROR at line 1:
ORA-01153: an incompatible media recovery is active
SQL> alter database recover managed standby database cancel;
Database altered.
SQL> alter database commit to switchover to primary;
Database altered. 转换成功之后，数据库状态为 nomount
SQL> select status from v$instance;
STATUS
------------------------
STARTED
Shut immediate
Statrtup
查看状态
SQL> select switchover_status from v$database;
SWITCHOVER_STATUS
----------------------------------------
SESSIONS ACTIVE
此时已经切换成功。如果真正的成功没有，那我们还需要做 日志传输和应用实验。
证明我们的switchover 成功了。
FAILOVER（没有做测试）
1、原则上在主库上不进行操作
因为既然是FAILOVER ,生产情况主库已经崩溃或者无法打开！！！无法执行，这里仅测试。
1、从库取消恢复
SQL> ALTER DATABASE RECOVER MANAGED STANDBY DATABASE FINISH FORCE;
---force关键字将会停止当前活动RFS进程，以便立即执行failover
2、备库直接切换为主库
SQL> ALTER DATABASE COMMIT TO SWITCHOVER TO PRIMARY;
3、启动数据库
SQL> SHUTDOWN IMMEDIATE;
SQL> STARTUP;
切换成功后原来主库已经不是dg里的一部分了。
接下来的工作就是将原来的主库恢复数据 或者重新改造成dg里的standby库！！！