Oracle 10g Data Guard Broker 配置文档

最新推荐文章于 2021-04-07 20:33:11 发布

小宝老豆

最新推荐文章于 2021-04-07 20:33:11 发布

阅读量883

点赞数

分类专栏： Oracle Data Guard 文章标签： oracle database archive system sql 数据库

Oracle Data Guard 专栏收录该内容

34 篇文章 0 订阅

订阅专栏

Oracle Data Guard Broker的配置及一些说明，参考Oracle 官网文档：

Oracle Data Guard 快速启动故障切换指南

http://www.oracle.com/technology/global/cn/pub/articles/smiley-fsfo.html

测试平台： Redhat 5.4 + Oracle 10gR2+ Physical Standby

一.  Data Guard  Broker 相关概念

1.1 Broker

Data Guard环境设置了Broker后，当主库出现一下情况，Broker就会自动进行主备库的切换。

1) Instance Failure

2) Shutdown Abort

3) Offline Datafiles due to I/O error

4) Network disconnection

切换时，Observer会将备库变成主库。同时Observer 通过应用Flashback log将原来的主库变成备库。所以，Data Guard Broker必须开启Flashback，同时将DG 配置在maximum availability模式。

Data Guard Broker 会在所有的主备库上使用自己的DMON进程和配置文件进行通信。当我们将DG_BROKER_START 设置为true之后，DMON进程就会启动。对于Broker的配置文件，我们可以通过Enterprise Manger 或者使用DGMGRL命令进行创建。

Broker的配置文件有自己的参数和设置。 Data Guard 允许有多分配置文件，所以我们需要在DG_BROKER_CONFIG_FILEn 参数里指定正确的配置文件。要注意的是，这个控制文件是二进制文件，不要手动去修改它，这样会破坏这个文件。

如果是RAC 环境，所有的实例有一份Broker 配置文件。该文件放在共享设备上。

Broker的更多内容参考：

Interaction Between the Data Guard Broker and a Data Guard Configuration [ID 249703.1]

    http://blog.csdn.net/xujinyang/article/details/6925015

1.2 Maximum Availability模式

最高可用性（Maximum availability）：这种模式在不影响Primary数据库可用前提下，提供最高级别的数据保护策略。其实现方式与最大保护模式类似，也是要求本地事务在提交前必须至少写入一台Standby数据库的Standby Redologs中，不过与最大保护模式不同的是，如果出现故障导致Standby数据库无法访问，Primary数据库并不会被Shutdown，而是自动转为最高性能模式，等Standby数据库恢复正常之后，Primary数据库又会自动转换成最高可用性模式。

Maximum protection/AVAILABILITY模式必须满足以下条件：

（1）Redo Archival Process: LGWR

（2）Network Tranmission mode: SYNC

（3）Disk Write Option: AFFIRM

（4）Standby Redo Logs: Yes

（5）standby database type: Physical Only

即Standby Database 必须配置Standby Redo Log，而Primary Database必须使用LGWR，SYNC，AFFIRM 方式归档到Standby Database.

如：

SQL> alter system set log_archive_dest_2='service=orcl_st lgwr sync AFFIRM';

注意：主库的保护模式修改之后，备库的模式也会改变，和主库保持一致。

在此对LGWR 进程的SYNC 方式做下说明：

（1）Primary Database 产生的Redo 日志要同时写道日志文件和网络。也就是说LGWR进程把日志写到本地日志文件的同时还要发送给本地的LNSn进程（Network Server Process），再由LNSn（LGWR Network Server process）进程把日志通过网络发送给远程的目的地，每个远程目的地对应一个LNS进程，多个LNS进程能够并行工作。

（2）LGWR 必须等待写入本地日志文件操作和通过LNSn进程的网络传送都成功，Primary Database 上的事务才能提交，这也是SYNC的含义所在。

（3）Standby Database的RFS进程把接收到的日志写入到Standby Redo Log日志中。

（4） Primary Database的日志切换也会触发Standby Database 上的日志切换，即Standby Database 对Standby Redo Log的归档，然后触发Standby Database 的MRP或者LSP 进程恢复归档日志。

因为Primary Database 的Redo 是实时传递的，于是Standby Database 端可以使用两种恢复方法：

实时恢复（Real-Time Apply）：只要RFS把日志写入Standby Redo Log 就会立即进行恢复；

归档恢复：在完成对Standby Redo Log 归档才触发恢复。

Primary Database默认使用ARCH进程，如果使用LGWR进程必须明确指定。使用LGWR SYNC方式时，可以同时使用NET_TIMEOUT参数，这个参数单位是秒，代表如果多长时间内网络发送没有响应，LGWR 进程会抛出错误。

示例：

alter system set log_archive_dest_2 = 'SERVICE=ST LGWR SYNC NET_TIMEOUT=30' scope=both;

Oracle Data Guard 理论知识，更多内容参考：

http://blog.csdn.net/xujinyang/article/details/6833263

二. 用DGMGRL配置物理standby的Broker示例

Oracle 10g官网文档参考：

IMPLEMENTING FAST-START FAILOVER IN 10GR2 DATAGUARD BROKER ENVIRONMENT [ID 359555.1]

http://blog.csdn.net/xujinyang/article/details/6833180

Oracle 9i 官网文档参考：

Setup and maintenance of Data Guard Broker using DGMGRL [ID 201669.1]

http://blog.csdn.net/xujinyang/article/details/6925035

这篇文章主要介绍Broker的配置，关于Data Guard的配置参考：

Oracle Data Guard Linux 平台 Physical Standby 搭建实例

http://blog.csdn.net/xujinyang/article/details/6829555

Oracle Data Guard Linux 平台 Logical Standby 创建实例

http://blog.csdn.net/xujinyang/article/details/6829546

在这里补充一点，我以前配置DG，喜欢使用默认方式，这样配置是非常简单的。但是最好还是把所有参数写全，其实也没有几个参数需要配置。关于参数的详细参考：

Oracle Data Guard 理论知识

http://blog.csdn.net/xujinyang/article/details/6833263

在pfile里添加如下参数，用pfile启动后，在创建spfile：

####主库参数######

*.DB_NAME ='orcl';

*.DB_UNIQUE_NAME='orcl_pd' --监听中配置的

*.LOG_ARCHIVE_CONFIG='DG_CONFIG=(orcl_pd,orcl_st)'

--列出DG中所有DB_UNIQUE_NAME

*.LOG_ARCHIVE_DEST_1='LOCATION=/u01/archive VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=orcl_pd'

*.LOG_ARCHIVE_DEST_2='service=orcl_st LGWR SYNC AFFIRM NET_TIMEOUT=30 DB_UNIQUE_NAME=orcl_st'

*.LOG_ARCHIVE_DEST_STATE_1=ENABLE

*.LOG_ARCHIVE_DEST_STATE_2=ENABLE

*.REMOTE_LOGIN_PASSWORDFILE=EXCLUSIVE

#####备库参数#######

*.FAL_SERVER=orcl_st

*.FAL_CLIENT=orcl_pd

*.standby_file_management='AUTO'

*.standby_archive_dest='/u01/archive'

#如果主备库目录不同，还需要添加：

*.log_file_name_convert='/u02/oradata/orcl/','/u03/oradata/orcl/'

*.db_file_name_convert='/u02/oradata/orcl/','/u03/oradata/orcl/'

在主备库的pfile里把如下参数都添加进去，因为切换之后还是需要使用的。不过主备库相应的目录和实例需要调整。

注意：必须配置local_listener，由各实例参数local_listener解析出来的监听地址必须能被所有成员访问 local_listener

主库： *. local_listener='orcl_pd;

备库： *. local_listener='orcl_st;

2.1 切换Data Guard到Maximum availability模式

默认情况下是Maximum performance，所以我们需要调整。

查看主备库log_archive_dest_2 参数：

（1）主库：

SQL> show parameter log_archive_dest_2

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

log_archive_dest_2 string SERVICE=orcl_st

（2）备库：

SQL> show parameter log_archive_dest_2

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

log_archive_dest_2 string SERVICE=orcl_pd

修改主备库的log_archive_dest_2参数：

（1）主库：

SQL> alter system set log_archive_dest_2='service=orcl_st LGWR SYNC AFFIRM NET_TIMEOUT=30 DB_UNIQUE_NAME=orcl_st';

System altered.

SQL> show parameter log_archive_dest_2

（2）备库：

SQL> alter system set log_archive_dest_2 = 'SERVICE=orcl_pd LGWR SYNC NET_TIMEOUT=30 AFFIRM DB_UNIQUE_NAME=orcl_pd' scope=both;

System altered.

SQL> show parameter log_archive_dest_2

在主备库查看模式：

SQL> select protection_mode,protection_level from v$database;

PROTECTION_MODE PROTECTION_LEVEL

-------------------- --------------------

MAXIMUM PERFORMANCE MAXIMUM PERFORMANCE

主备库都执行：

注意，切换模式要在非open 状态执行，所以主库需要shutdown，在启动到mount后，在执行命令，备库直接执行即可。

SQL> alter database set standby database to maximize availability;

Database altered.

SQL> select protection_mode,protection_level from v$database;

PROTECTION_MODE PROTECTION_LEVEL

-------------------- --------------------

MAXIMUM AVAILABILITY MAXIMUM AVAILABILITY

2.2 启动Flashback

数据库的Flashback Database功能缺省是关闭的。我们需要启动它。

可以通过如下SQL 查看：

SQL> select flashback_on from v$database;

FLASHBACK_ON

------------------

在启用Flashback database 之前，我们需要先设定闪回区：

在主备库查看：

SQL> show parameter db_recovery_file_dest

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

db_recovery_file_dest string /u01/app/oracle/flash_recovery_area

db_recovery_file_dest_size big integer 2G

我这里是设置过了，如果没有设置，可以用以下语句进行设置：

Alter system set db_recovery_file_dest_size=<integer>[<k><M><G>];

Alter system set db_recovery_file_dest=<path_to_flashabck_logs>;

启动flashback database.

主库：这个操作也需要在非open 状态执行。主库需要shutdown后，启动到mount 状态执行。

SQL> Alter database flashback on;

备库：需要先取消recover 进程，不然会报错。

SQL> Alter database flashback on;

Alter database flashback on

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 flashback on;

Database altered.

SQL> select flashback_on from v$database;

FLASHBACK_ON

------------------

YES

SQL> alter database recover managed standby database disconnect from session;

Database altered.

详细内容参考：

Oracle Flashback 技术总结

http://blog.csdn.net/xujinyang/article/details/6830438

2.3 安装DGMGRL

这一步是官网上建议安装的。在另一台观察机器observer 上安装DGMGRL。就是安装一个Oracle 的客户端。并在observer machine上配置相关的参数。包括配置监听，使这台机器能访问主备库的实例。然后通过这个observer来判断主备库的状态。如果主库出现问了，那么observer 就会切换主备库。

放在另一台机器的原因也很明显，如果放在主库上，如果主库系统崩溃了，那么Observer也就失效了。

在这个实验中，我不在单独在其他机器上安装DGMGRL命令。我直接在备库的系统上配置这个observer.

2.4 设置broker 参数

在主备库查看dg_broker_start参数：

SQL> show parameter dg_broker_start

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

dg_broker_start boolean FALSE

SQL> alter system set dg_broker_start = true scope=both;

System altered.

SQL> show parameter dg_broker_start

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

dg_broker_start boolean TRUE

SQL>

2.5 配置Listener.ora 文件

To enable DGMGRL to restart instances during the course of broker operations, a service with a specific name must be statically registered with the local listener of each instance. A static service registration is also required to enable the observer to restart instances as part of automatic reinstatement of the old primary database after a fast-start failover has occurred. The broker uses a default name for the GLOBAL_DBNAME attribute of db_unique_name_DGMGRL.db_domain.

From：http://download.oracle.com/docs/cd/E18283_01/server.112/e17023/install.htm

主备库的Listener.ora 文件都修改成如下格式：

[oracle@dg1 admin]$ cat listener.ora

SID_LIST_LISTENER =

(SID_LIST =

(SID_DESC =

(SID_NAME = PLSExtProc)

(ORACLE_HOME = /u01/app/oracle/product/10.2.0/db_1)

(PROGRAM = extproc)

)

(SID_DESC =

(SID_NAME = orcl)

(ORACLE_HOME = /u01/app/oracle/product/10.2.0/db_1)

(GLOBAL_DBNAME = orcl)

)

(SID_DESC =

(GLOBAL_DBNAME = orcl_pd_DGMGRL)

(ORACLE_HOME = /u01/app/oracle/product/10.2.0/db_1)

(SID_NAME = orcl)

)

注意这里的GLOBAL_DBNAME参数格式：

<db_unique_name>_DGMGRL.<db_domain> 的连接。

备库改成备库的db_unique_name 就可以了。并且db_unique_name 大小写敏感。其他一样。

或者使用net manager工具，在listeners选项下，选择Database Services服务，添加一下。

最好不要忘了用: lsnrctl reload 命令重新加载一下。

2.6 设置FAST_START FAILOVER

这里是测试，不在用另一台机器来做observer。故在备库上配置Broker.

[oracle@dg2 /]$ dgmgrl

DGMGRL for Linux: Version 10.2.0.1.0 - Production

Welcome to DGMGRL, type "help" for information.

DGMGRL> connect sys/oracle@orcl_pd;

Connected.

DGMGRL> help create

Create a broker configuration

Syntax:

CREATE CONFIGURATION <configuration name> AS

PRIMARY DATABASE IS <database name> -- db_unique_name

CONNECT IDENTIFIER IS <connect identifier>; -- 监听中配置的名称

DGMGRL> create configuration 'OrclBroker' as primary database is 'orcl_pd' connect identifier is orcl_pd;

Configuration "OrclBroker" created with primary database "orcl_pd"

DGMGRL> help add

Add a standby database to the broker configuration

Syntax:

ADD DATABASE <database name> AS

CONNECT IDENTIFIER IS <connect identifier>

MAINTAINED AS {PHYSICAL|LOGICAL};

DGMGRL> add database 'orcl_st' as connect identifier is orcl_st maintained as physical;

Database "orcl_st" added

DGMGRL> show configuration

Configuration

Name: OrclBroker

Enabled: NO

Protection Mode: MaxAvailability

Fast-Start Failover: DISABLED

Databases:

orcl_pd - Primary database

orcl_st - Physical standby database

Current status for "OrclBroker":

DISABLED

如果出现这种错误，处理方法参考：

ORA-16796: one or more properties could not be imported from the database [ID 358040.1]

http://blog.csdn.net/xujinyang/article/details/6833155

现在会在主库的$ORACLE_HOME/dbs/下生成2个配置文件：dr1orcl_pd.dat 和dr2orcl_pd.dat. 这2个文件的位置有如下参数指定：

SQL> show parameter dg_broker_config_file

NAME TYPE VALUE

------------------------------------ ----------- ------------------------------

dg_broker_config_file1 string /u01/app/oracle/product/10.2.0

/db_1/dbs/dr1orcl_pd.dat

dg_broker_config_file2 string /u01/app/oracle/product/10.2.0

/db_1/dbs/dr2orcl_pd.dat

[oracle@dg1 dbs]$ pwd

/u01/app/oracle/product/10.2.0/db_1/dbs

[oracle@dg1 dbs]$ ls

dr1orcl_pd.dat hc_orcl.dat init.ora lkORCL orapworcl

dr2orcl_pd.dat initdw.ora initorcl.ora lkORCL_PD spfileorcl.ora

DGMGRL 的日志位置在：

[oracle@dg2 admin]$ cd /u01/app/oracle/admin/orcl/bdump/

[oracle@dg2 bdump]$ ls

alert_orcl.log drcorcl.log

[oracle@dg2 bdump]$ cat dr

现在我们启用configuration：

DGMGRL> enable configuration

Enabled.

启用之后，在备库的相关目录下也会生成这2个文件。这2个文件在启用之前是没有生成的。

DGMGRL> show configuration

Configuration

Name: OrclBroker

Enabled: YES

Protection Mode: MaxAvailability

Fast-Start Failover: DISABLED --FSFO 是禁用的

Databases:

orcl_pd - Primary database

orcl_st - Physical standby database

Current status for "OrclBroker":

SUCCESS

DGMGRL> show database verbose 'orcl_pd';

Database

Name: orcl_pd

Role: PRIMARY

Enabled: YES

Intended State: ONLINE

Instance(s):

orcl

Properties:

InitialConnectIdentifier = 'orcl_pd'

LogXptMode = 'ASYNC' --状态不正确，等会调整

Dependency = ''

DelayMins = '0'

Binding = 'OPTIONAL'

MaxFailure = '0'

MaxConnections = '1'

ReopenSecs = '300'

NetTimeout = '180'

LogShipping = 'ON'

PreferredApplyInstance = ''

ApplyInstanceTimeout = '0'

ApplyParallel = 'AUTO'

StandbyFileManagement = 'AUTO'

ArchiveLagTarget = '0'

LogArchiveMaxProcesses = '2'

LogArchiveMinSucceedDest = '1'

DbFileNameConvert = ''

LogFileNameConvert = ''

FastStartFailoverTarget = ''

StatusReport = '(monitor)'

InconsistentProperties = '(monitor)'

InconsistentLogXptProps = '(monitor)'

SendQEntries = '(monitor)'

LogXptStatus = '(monitor)'

RecvQEntries = '(monitor)'

HostName = 'dg1'

SidName = 'orcl'

LocalListenerAddress = '(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.2)(PORT=1521))'

StandbyArchiveLocation = '/u01/archive'

AlternateLocation = ''

LogArchiveTrace = '0'

LogArchiveFormat = '%t_%s_%r.dbf'

LatestLog = '(monitor)'

TopWaitEvents = '(monitor)'

Current status for "orcl_pd":

SUCCESS

DGMGRL> show database verbose 'orcl_st'

Database

Name: orcl_st

Role: PHYSICAL STANDBY

Enabled: YES

Intended State: ONLINE

Instance(s):

orcl

Properties:

InitialConnectIdentifier = 'orcl_st'

LogXptMode = 'SYNC'

Dependency = ''

DelayMins = '0'

Binding = 'OPTIONAL'

MaxFailure = '0'

MaxConnections = '1'

ReopenSecs = '300'

NetTimeout = '30'

LogShipping = 'ON'

PreferredApplyInstance = ''

ApplyInstanceTimeout = '0'

ApplyParallel = 'AUTO'

StandbyFileManagement = 'AUTO'

ArchiveLagTarget = '0'

LogArchiveMaxProcesses = '2'

LogArchiveMinSucceedDest = '1'

DbFileNameConvert = ''

LogFileNameConvert = ''

FastStartFailoverTarget = ''

StatusReport = '(monitor)'

InconsistentProperties = '(monitor)'

InconsistentLogXptProps = '(monitor)'

SendQEntries = '(monitor)'

LogXptStatus = '(monitor)'

RecvQEntries = '(monitor)'

HostName = 'dg2'

SidName = 'orcl'

LocalListenerAddress = '(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.3)(PORT=1521))'

StandbyArchiveLocation = '/u01/archive'

AlternateLocation = ''

LogArchiveTrace = '0'

LogArchiveFormat = '%t_%s_%r.dbf'

LatestLog = '(monitor)'

TopWaitEvents = '(monitor)'

Current status for "orcl_st":

SUCCESS

在enable FSFO 之前，需要注意LogXptMode的状态，必须为SYNC。刚才看到orcl_pd 的状态不正确，我们调整一下：

DGMGRL> edit database orcl_pd set property logxptmode=sync;

Property "logxptmode" updated

DGMGRL> show database verbose orcl_pd

Database

Name: orcl_pd

Role: PRIMARY

Enabled: YES

Intended State: ONLINE

Instance(s):

orcl

Properties:

InitialConnectIdentifier = 'orcl'

LogXptMode = 'sync'

Dependency = ''

...

不然在启用FSFO时，会遇到如下错误：

ORA-16651: requirements not met for enabling Fast-Start Failover

ORA-16651:

requirements not met for enabling Fast-Start Failover

Cause:

The attempt to enable Fast-Start Failover could not be completed because one or more requirements have not been met:

- The Data Guard configuration must be in MaxAvailability protection mode.

- The LogXptMode property for both the primary database and the Fast-Start Failover target standby database must be SYNC.

- The primary database and the Fast-Start Failover target standby database must both have flashback enabled.

- No valid target standby database was specified in the primary database's FastStartFailoverTarget property prior to the attempt to enable Fast-Start Failover, and more than one standby database exists in the Data Guard configuration.

Action:

Retry the attempted command after correcting the issue:

- Set the Data Guard configuration to MaxAvailability protection mode.

- Ensure that the LogXptMode property for both the primary database and the Fast-Start Failover target standby database are SYNC.

- Ensure that both the primary database and the Fast-Start Failover target standby database have flashback enabled.

- Set the primary database's FastStartFailoverTarget property to the db_unique_name value of the desired target standby database add the desired target standby database's FastStartFailoverTarget property to the db_unique_name value of the primary database.

启用FSFO：

DGMGRL> enable fast_start failover;

Enabled.

DGMGRL> show configuration verbose

Configuration

Name: OrclBroker

Enabled: YES

Protection Mode: MaxAvailability

Fast-Start Failover: ENABLED

Databases:

orcl_pd - Primary database

orcl_st - Physical standby database

- Fast-Start Failover target

Fast-Start Failover

Threshold: 30 seconds --默认值太小，等会改成120s

Observer: (none)

Current status for "OrclBroker":

Warning: ORA-16608: one or more databases have warnings -- 警告因为Observer没有启动，这个从DGMGRL 日志中可以看到。

DGMGRL> edit configuration set property FastStartFailoverThreshold=120;

Property "faststartfailoverthreshold" updated

DGMGRL> show database verbose orcl_pd

Database

Name: orcl_pd

Role: PRIMARY

Enabled: YES

Intended State: ONLINE

Instance(s):

orcl

Properties:

InitialConnectIdentifier = 'orcl'

LogXptMode = 'sync'

Dependency = ''

DelayMins = '0'

Binding = 'OPTIONAL'

MaxFailure = '0'

MaxConnections = '1'

ReopenSecs = '300'

NetTimeout = '180'

LogShipping = 'ON'

PreferredApplyInstance = ''

ApplyInstanceTimeout = '0'

ApplyParallel = 'AUTO'

StandbyFileManagement = 'AUTO'

ArchiveLagTarget = '0'

LogArchiveMaxProcesses = '2'

LogArchiveMinSucceedDest = '1'

DbFileNameConvert = ''

LogFileNameConvert = ''

FastStartFailoverTarget = 'orcl_st'

StatusReport = '(monitor)'

InconsistentProperties = '(monitor)'

InconsistentLogXptProps = '(monitor)'

SendQEntries = '(monitor)'

LogXptStatus = '(monitor)'

RecvQEntries = '(monitor)'

HostName = 'dg1'

SidName = 'orcl'

LocalListenerAddress = '(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.2)(PORT=1521))'

StandbyArchiveLocation = '/u01/archive'

AlternateLocation = ''

LogArchiveTrace = '0'

LogArchiveFormat = '%t_%s_%r.dbf'

LatestLog = '(monitor)'

TopWaitEvents = '(monitor)'

Current status for "orcl_pd":

Warning: ORA-16819: Fast-Start Failover observer not started -- 从这里也可以清楚的看出问题

启动Observer

DGMGRL> start observer

Observer started

注意，启动之后，该前台进程不会退出，会一直挂在这。直到从其他窗口关闭。

默认情况下，observer会创建一个二进制的文件 fsfo.dat来保存主库和备库的连接信息。这个文件会在调用dgmgrl命令的当前窗口下生成。

2.7 验证

主库：

SQL> select fs_failover_observer_present,fs_failover_observer_host,fs_failover_threshold from v$database;

FS_FAIL FS_FAILOVER_OBSERVER FS_FAILOVER_THRESHOLD

------- -------------------- ---------------------

YES dg2 120

备库：

SQL> select fs_failover_observer_present,fs_failover_observer_host,fs_failover_threshold from v$database;

FS_FAIL FS_FAILOVER_OBSERVER FS_FAILOVER_THRESHOLD

------- -------------------- ---------------------

YES dg2 120

[oracle@dg2 dbs]$ dgmgrl

DGMGRL for Linux: Version 10.2.0.1.0 - Production

Welcome to DGMGRL, type "help" for information.

DGMGRL> connect sys/oracle@orcl_pd;

Connected.

DGMGRL> show configuration verbose;

Configuration

Name: OrclBroker

Enabled: YES

Protection Mode: MaxAvailability

Fast-Start Failover: ENABLED

Databases:

orcl_pd - Primary database

orcl_st - Physical standby database

- Fast-Start Failover target

Fast-Start Failover

Threshold: 120 seconds

Observer: dg2

Current status for "OrclBroker":

SUCCESS

DGMGRL> show database verbose orcl_pd;

Database

Name: orcl_pd

Role: PRIMARY

Enabled: YES

Intended State: ONLINE

Instance(s):

orcl

Properties:

InitialConnectIdentifier = 'orcl'

LogXptMode = 'sync'

Dependency = ''

DelayMins = '0'

Binding = 'OPTIONAL'

MaxFailure = '0'

MaxConnections = '1'

ReopenSecs = '300'

NetTimeout = '180'

LogShipping = 'ON'

PreferredApplyInstance = ''

ApplyInstanceTimeout = '0'

ApplyParallel = 'AUTO'

StandbyFileManagement = 'AUTO'

ArchiveLagTarget = '0'

LogArchiveMaxProcesses = '2'

LogArchiveMinSucceedDest = '1'

DbFileNameConvert = ''

LogFileNameConvert = ''

FastStartFailoverTarget = 'orcl_st'

StatusReport = '(monitor)'

InconsistentProperties = '(monitor)'

InconsistentLogXptProps = '(monitor)'

SendQEntries = '(monitor)'

LogXptStatus = '(monitor)'

RecvQEntries = '(monitor)'

HostName = 'dg1'

SidName = 'orcl'

LocalListenerAddress = '(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.2)(PORT=1521))'

StandbyArchiveLocation = '/u01/archive'

AlternateLocation = ''

LogArchiveTrace = '0'

LogArchiveFormat = '%t_%s_%r.dbf'

LatestLog = '(monitor)'

TopWaitEvents = '(monitor)'

Current status for "orcl_pd":

SUCCESS

DGMGRL> show database verbose orcl_st;

Database

Name: orcl_st

Role: PHYSICAL STANDBY

Enabled: YES

Intended State: ONLINE

Instance(s):

orcl

Properties:

InitialConnectIdentifier = 'orcl_st'

LogXptMode = 'SYNC'

Dependency = ''

DelayMins = '0'

Binding = 'OPTIONAL'

MaxFailure = '0'

MaxConnections = '1'

ReopenSecs = '300'

NetTimeout = '30'

LogShipping = 'ON'

PreferredApplyInstance = ''

ApplyInstanceTimeout = '0'

ApplyParallel = 'AUTO'

StandbyFileManagement = 'AUTO'

ArchiveLagTarget = '0'

LogArchiveMaxProcesses = '2'

LogArchiveMinSucceedDest = '1'

DbFileNameConvert = ''

LogFileNameConvert = ''

FastStartFailoverTarget = 'orcl_pd'

StatusReport = '(monitor)'

InconsistentProperties = '(monitor)'

InconsistentLogXptProps = '(monitor)'

SendQEntries = '(monitor)'

LogXptStatus = '(monitor)'

RecvQEntries = '(monitor)'

HostName = 'dg2'

SidName = 'orcl'

LocalListenerAddress = '(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.3)(PORT=1521))'

StandbyArchiveLocation = '/u01/archive'

AlternateLocation = ''

LogArchiveTrace = '0'

LogArchiveFormat = '%t_%s_%r.dbf'

LatestLog = '(monitor)'

TopWaitEvents = '(monitor)'

Current status for "orcl_st":

SUCCESS

2.8 验证自动切换

在前面提到，在一下情况会发生切换：

1) Instance Failure

2) Shutdown Abort

3) Offline Datafiles due to I/O error

4) Network disconnection

所以我们这里模拟主库shutdown的情况，我们在主库执行shutdown abort 在查看主备库的情况。

（1）先在客户端配置一下TAF

在tnsnames.ora 文件里添加如下参数：

TAFTEST=

(DESCRIPTION=

(LOAD_BALANCE=on)

(FAILOVER=on)

(ADDRESS=

(PROTOCOL=tcp)

(HOST=192.168.3.2)

(PORT=1521))

(ADDRESS=

(PROTOCOL=tcp)

(HOST=192.168.3.3)

(PORT=1521))

(CONNECT_DATA=

(SERVICE_NAME=orcl)

(FAILOVER_MODE=

(TYPE=select)

(METHOD=basic))))

然后通过客户端连接这个是里，连接上后我们查看一下：

C:/Users/Administrator.DavidDai>sqlplus /nolog

SQL*Plus: Release 11.2.0.1.0 Production on 星期一 12月 13 16:56:33 2010

SQL> conn sys/oracle@taftest as sysdba;

已连接。

SQL> select db_unique_name from v$database;

DB_UNIQUE_NAME

------------------------------

orcl_pd

SQL> select name from v$database;

NAME

---------

ORCL

设置TAF之后，当Data Guard 切换之后，我们在客户端不需要做什么修改，还可以继续连接数据库。关于TAF 的内容，参考：

Oracle Transparent Application Failover(TAF) 说明

http://blog.csdn.net/xujinyang/article/details/6831349

（2）shutdown about 主库

SQL> shutdown abort

ORACLE instance shut down.

我这里是采用shutdown 来验证，当然也可以用switchover to orcl_st 来进行切换。这种做法可以参考：

Oracle Data Guard 快速启动故障切换指南

http://www.oracle.com/technology/global/cn/pub/articles/smiley-fsfo.html

（3）查看Broker log

2分钟之后开始切换，我们可以从主备库observer的log中查看：

原备库：

[oracle@dg2 bdump]$ pwd

/u01/app/oracle/admin/orcl/bdump

[oracle@dg2 bdump]$ tail -100 drcorcl.log

DG 2010-12-14-01:03:55 0 2 0 Database Resource[IAM=PRIMARY]: SetState READ-WRITE-XPTON, phase BUILD-UP, External Cond GRACEFUL-FAIL-OVER-PHYS_STBY, Target Site Handle 0x02001000

DG 2010-12-14-01:03:55 0 2 0 Executing SQL [ALTER SYSTEM SET standby_archive_dest='' SCOPE=BOTH sid='orcl']

DG 2010-12-14-01:03:55 0 2 0 SQL [ALTER SYSTEM SET standby_archive_dest='' SCOPE=BOTH sid='orcl'] Executed successfully

DG 2010-12-14-01:03:55 0 2 0 Executing SQL [ALTER SYSTEM SET log_archive_dest_1='location="/u01/archive"', 'valid_for=(ONLINE_LOGFILE,ALL_ROLES)' scope=both sid='orcl' ]

DG 2010-12-14-01:03:55 0 2 0 SQL [ALTER SYSTEM SET log_archive_dest_1='location="/u01/archive"', 'valid_for=(ONLINE_LOGFILE,ALL_ROLES)' scope=both sid='orcl' ] Executed successfully

DG 2010-12-14-01:03:55 0 2 0 Executing SQL [ALTER SYSTEM SET log_archive_dest_state_1='ENABLE' SCOPE=BOTH sid='orcl']

DG 2010-12-14-01:03:55 0 2 0 SQL [ALTER SYSTEM SET log_archive_dest_state_1='ENABLE' SCOPE=BOTH sid='orcl'] Executed successfully

DG 2010-12-14-01:03:55 0 2 0 Executing SQL [ALTER DATABASE SET STANDBY DATABASE TO MAXIMIZE AVAILABILITY]

DG 2010-12-14-01:03:55 0 2 0 SQL [ALTER DATABASE SET STANDBY DATABASE TO MAXIMIZE AVAILABILITY] Executed successfully

DG 2010-12-14-01:03:55 0 2 0 Set log transport destination: SetState ONLINE, phase BUILD-UP, External Cond GRACEFUL-FAIL-OVER-PHYS_STBY

DG 2010-12-14-01:03:55 0 2 0 Executing SQL [ALTER DATABASE OPEN]

已经完成了切换。

在客户端查看一下：

SQL> conn sys/oracle@taftest as sysdba;

已连接。

SQL> select instance_name from v$instance;

INSTANCE_NAME

----------------

orcl

SQL> select db_unique_name from v$database;

DB_UNIQUE_NAME

------------------------------

orcl_st

在原来observer 启动的地方也会有提示：

DGMGRL> start observer

Observer started

01:03:01.88 Tuesday, December 14, 2010

Initiating fast-start failover to database "orcl_st"...

Performing failover NOW, please wait...

Failover succeeded, new primary is "orcl_st"

01:04:53.38 Tuesday, December 14, 2010

原来的备库：

SQL> select open_mode,database_role from v$database;

OPEN_MODE DATABASE_ROLE

---------- ----------------

READ WRITE PRIMARY

SQL> select fs_failover_status from v$database;

FS_FAILOVER_STATUS

---------------------

REINSTATE REQUIRED -- 表明原来的主库还没有被切换为备库

用dgmgrl连接到备库看看：

[oracle@dg2 dbs]$ dgmgrl

DGMGRL for Linux: Version 10.2.0.1.0 - Production

Welcome to DGMGRL, type "help" for information.

DGMGRL> connect sys/oracle@orcl_st;

Connected.

DGMGRL> show database verbose orcl_st;

Database

Name: orcl_st

Role: PRIMARY

Enabled: YES

Intended State: ONLINE

Instance(s):

orcl

Properties:

InitialConnectIdentifier = 'orcl_st'

LogXptMode = 'SYNC'

Dependency = ''

DelayMins = '0'

Binding = 'OPTIONAL'

MaxFailure = '0'

MaxConnections = '1'

ReopenSecs = '300'

NetTimeout = '30'

LogShipping = 'ON'

PreferredApplyInstance = ''

ApplyInstanceTimeout = '0'

ApplyParallel = 'AUTO'

StandbyFileManagement = 'AUTO'

ArchiveLagTarget = '0'

LogArchiveMaxProcesses = '2'

LogArchiveMinSucceedDest = '1'

DbFileNameConvert = ''

LogFileNameConvert = ''

FastStartFailoverTarget = 'orcl_pd'

StatusReport = '(monitor)'

InconsistentProperties = '(monitor)'

InconsistentLogXptProps = '(monitor)'

SendQEntries = '(monitor)'

LogXptStatus = '(monitor)'

RecvQEntries = '(monitor)'

HostName = 'dg2'

SidName = 'orcl'

LocalListenerAddress = '(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.3)(PORT=1521))'

StandbyArchiveLocation = '/u01/archive'

AlternateLocation = ''

LogArchiveTrace = '0'

LogArchiveFormat = '%t_%s_%r.dbf'

LatestLog = '(monitor)'

TopWaitEvents = '(monitor)'

Current status for "orcl_st":

Warning: ORA-16817: unsynchronized Fast-Start Failover configuration

DGMGRL> show database verbose orcl_pd;

Database

Name: orcl_pd

Role: PHYSICAL STANDBY

Enabled: NO

Intended State: ONLINE

Instance(s):

orcl

Properties:

InitialConnectIdentifier = 'orcl'

LogXptMode = 'sync'

Dependency = ''

DelayMins = '0'

Binding = 'OPTIONAL'

MaxFailure = '0'

MaxConnections = '1'

ReopenSecs = '300'

NetTimeout = '180'

LogShipping = 'ON'

PreferredApplyInstance = ''

ApplyInstanceTimeout = '0'

ApplyParallel = 'AUTO'

StandbyFileManagement = 'AUTO'

ArchiveLagTarget = '0'

LogArchiveMaxProcesses = '2'

LogArchiveMinSucceedDest = '1'

DbFileNameConvert = ''

LogFileNameConvert = ''

FastStartFailoverTarget = 'orcl_st'

StatusReport = '(monitor)'

InconsistentProperties = '(monitor)'

InconsistentLogXptProps = '(monitor)'

SendQEntries = '(monitor)'

LogXptStatus = '(monitor)'

RecvQEntries = '(monitor)'

HostName = 'dg1'

SidName = 'orcl'

LocalListenerAddress = '(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.2)(PORT=1521))'

StandbyArchiveLocation = 'dgsby_orcl_pd'

AlternateLocation = ''

LogArchiveTrace = '0'

LogArchiveFormat = '%t_%s_%r.dbf'

LatestLog = '(monitor)'

TopWaitEvents = '(monitor)'

Current status for "orcl_pd":

Error: ORA-16661: the standby database needs to be reinstated

（3）将原主库启动到startup mount状态

Broker 需要 SYSDBA 凭证以重新启动两个数据库或其中一个数据库。如果没有该凭证，Broker 仍将完成角色转换，但需要手动重新启动数据库.

将原主库启动到mount状态之后，Observer 会使用当前主库的Flashback Logs，将原来的备库切换到mount状态状态。

恢复的第一步是将数据库闪回到备用数据库变为主数据库的 SCN 处（新主数据库上的 v$database.standby_became_primary_scn）。闪回数据库将分成两个阶段进行：恢复阶段和介质恢复阶段。

在恢复阶段，闪回数据库使用闪回数据库日志中的前映像块将数据库恢复到 standby_became_primary_scn 之前的一点。在介质恢复阶段中，闪回数据库应用重做以将数据库带到standby_became_primary_scn。为使闪回数据库成功，闪回数据库日志中必须包括足够的可用历史记录，并且恢复点和 standby_became_primary_scn 之间生成的所有重做必须可用。如果闪回数据库失败，自动恢复将停止，您将需要手动执行基于 SCN 的恢复以恢复到 standby_became_primary_scn，直到完成该恢复。

一旦闪回数据库成功，观察器会将该数据库转换为备用数据库，执行回弹并开始应用服务。

原主库的log会有如下信息：

[oracle@dg1 bdump]$ D

[oracle@dg1 bdump]$ pwd

/u01/app/oracle/admin/orcl/bdump

[oracle@dg1 bdump]$ ls

alert_orcl.log orcl_arc0_5217.trc orcl_mrp0_5237.trc orcl_nsv1_5173.trc orcl_rsm0_5177.trc

drcorcl.log orcl_arc1_5219.trc orcl_nsv1_5165.trc orcl_pmon_5131.trc orcl_rsm0_5233.trc

[oracle@dg1 bdump]$ tail -100 drcorcl.log

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM REGISTER]

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM REGISTER] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET log_archive_dest_1='location="/u01/archive"', 'valid_for=(ALL_LOGFILES,ALL_ROLES)' scope=both sid='orcl' ]

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET log_archive_dest_1='location="/u01/archive"', 'valid_for=(ALL_LOGFILES,ALL_ROLES)' scope=both sid='orcl' ] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET log_archive_dest_state_1='ENABLE' SCOPE=BOTH sid='orcl']

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET log_archive_dest_state_1='ENABLE' SCOPE=BOTH sid='orcl'] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET standby_archive_dest='/u01/archive' SCOPE=BOTH sid='orcl']

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET standby_archive_dest='/u01/archive' SCOPE=BOTH sid='orcl'] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Setting init.ora parameter with SQL [ALTER SYSTEM SET log_archive_trace=0 SCOPE=BOTH sid='orcl']

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET log_archive_trace=0 SCOPE=BOTH sid='orcl']

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET log_archive_trace=0 SCOPE=BOTH sid='orcl'] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Setting init.ora parameter with SQL [ALTER SYSTEM SET log_archive_format='%t_%s_%r.dbf' SCOPE=SPFILE sid='orcl']

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET log_archive_format='%t_%s_%r.dbf' SCOPE=SPFILE sid='orcl']

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET log_archive_format='%t_%s_%r.dbf' SCOPE=SPFILE sid='orcl'] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Setting init.ora parameter with SQL [ALTER SYSTEM SET standby_file_management='AUTO' SCOPE=BOTH sid='*']

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET standby_file_management='AUTO' SCOPE=BOTH sid='*']

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET standby_file_management='AUTO' SCOPE=BOTH sid='*'] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Setting init.ora parameter with SQL [ALTER SYSTEM SET archive_lag_target=0 SCOPE=BOTH sid='*']

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET archive_lag_target=0 SCOPE=BOTH sid='*']

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET archive_lag_target=0 SCOPE=BOTH sid='*'] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Setting init.ora parameter with SQL [ALTER SYSTEM SET log_archive_max_processes=2 SCOPE=BOTH sid='*']

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET log_archive_max_processes=2 SCOPE=BOTH sid='*']

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET log_archive_max_processes=2 SCOPE=BOTH sid='*'] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Setting init.ora parameter with SQL [ALTER SYSTEM SET log_archive_min_succeed_dest=1 SCOPE=BOTH sid='*']

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET log_archive_min_succeed_dest=1 SCOPE=BOTH sid='*']

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET log_archive_min_succeed_dest=1 SCOPE=BOTH sid='*'] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET fal_server='(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.3)(PORT=1521)))(CONNECT_DATA=(SERVICE_NAME=orcl_st_XPT)(SERVER=dedicated)))']

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET fal_server='(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.3)(PORT=1521)))(CONNECT_DATA=(SERVICE_NAME=orcl_st_XPT)(SERVER=dedicated)))'] Executed successfully

DG 2010-12-14-01:12:12 0 2 0 Executing SQL [ALTER SYSTEM SET fal_client='(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.2)(PORT=1521)))(CONNECT_DATA=(SERVICE_NAME=orcl_pd_XPT)(INSTANCE_NAME=orcl)(SERVER=dedicated)))']

DG 2010-12-14-01:12:12 0 2 0 SQL [ALTER SYSTEM SET fal_client='(DESCRIPTION=(ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=192.168.3.2)(PORT=1521)))(CONNECT_DATA=(SERVICE_NAME=orcl_pd_XPT)(INSTANCE_NAME=orcl)(SERVER=dedicated)))'] Executed successfully

（4）最后验证切换之后的状态

在原备库：

SQL> select open_mode,database_role from v$database;

OPEN_MODE DATABASE_ROLE

---------- ----------------

READ WRITE PRIMARY

SQL> select fs_failover_status from v$database;

FS_FAILOVER_STATUS

---------------------

SYNCHRONIZED

原主库：

SQL> connect sys/oracle as sysdba;