DG 日志级联传送 参数据配置

DG级联安装：
数据库版本 11.2.0.1
r2  primary     10.3.4.110   
r5  standby1    10.3.4.111
r6  standby2    10.3.4.112

###########################################################主库r2参数文件设置#########################################
现在r2与r5是一套DG
主库r2的参数文件如下：
*.DB_UNIQUE_NAME='r2'
*.FAL_CLIENT='r2'
*.FAL_SERVER='r5'
*.LOG_ARCHIVE_CONFIG='DG_CONFIG=(r2,r5)'
*.LOG_ARCHIVE_DEST_1='LOCATION=/u01/archivelog  VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=r2'
*.LOG_ARCHIVE_DEST_2='SERVICE=r5 LGWR SYNC AFFIRM VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=r5'
*.LOG_ARCHIVE_DEST_STATE_1='ENABLE'
*.LOG_ARCHIVE_DEST_STATE_2='DEFER'
*.STANDBY_FILE_MANAGEMENT='AUTO'
*.DB_FILE_NAME_CONVERT='r2','r5'
*.LOG_FILE_NAME_CONVERT='r2','r5'

对主库的参数文件作修改如下：
ALTER SYSTEM SET LOG_ARCHIVE_DEST_3='SERVICE=dg2 LGWR  ASYNC AFFIRM VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=dg2' scope=both;
ALTER SYSTEM SET LOG_ARCHIVE_CONFIG='DG_CONFIG=(r2,r5,r6)' scope=both;
ALTER SYSTEM SET LOG_ARCHIVE_CONFIG='DEFER' scope=both;
#是为了后面的从主库r2的备库塔建备库r6
ALTER SYSTEM SET DB_FILE_NAME_CONVERT='r2','r6'  ;
ALTER SYSTEM SET LOG_FILE_NAME_CONVERT='r2','r6' ;


修改之后主库r2的参数文件如下：
*.DB_UNIQUE_NAME='r2'
*.FAL_CLIENT='r2'
*.FAL_SERVER='r5'
*.LOG_ARCHIVE_CONFIG='DG_CONFIG=(r2,r5,r6)'
*.LOG_ARCHIVE_DEST_1='LOCATION=/u01/archivelog  VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=r2'
*.LOG_ARCHIVE_DEST_2='SERVICE=r5 LGWR SYNC AFFIRM VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=r5'
*.LOG_ARCHIVE_DEST_2='SERVICE=r6 LGWR SYNC AFFIRM VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=r6'
*.LOG_ARCHIVE_DEST_STATE_1='ENABLE'
*.LOG_ARCHIVE_DEST_STATE_2='ENABLE'
*.LOG_ARCHIVE_DEST_STATE_2='DEFER'
*.STANDBY_FILE_MANAGEMENT='AUTO'
*.DB_FILE_NAME_CONVERT='r2','r6'
*.LOG_FILE_NAME_CONVERT='r2','r6'


###########################################  塔建备库r6 ###############################################################

在主库上复制备分到备库r6上,源数据库必须是主库

复制主库r2的参数文件至r6,并启动备库r6至mount状态

主库r2上运行：
rman target sys/admin@r2 auxiliary sys/admin@r6
duplicate target database for standby nofilenamecheck from active database;

alter system set log_archive_dest_state_3=enable scope=both;

在备库r6
alter database open read only;  
alter database recover managed standby database using current logfile disconnect from session;  
ALTER SYSTEM SET log_archive_dest_1='LOCATION=/u01/ora_archive VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=r6' scope=both;
ALTER SYSTEM SET log_archive_dest_2='SERVICE=r5 LGWR SYNC AFFIRM VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=r5'
alter system set fal_server='r5';  
alter system set fal_client='r6';

在主库r2上
alter system set log_archive_dest_state_3=defer scope=both;


###########################################  备库r5 参数的修改 #########################################################
*.fal_client='r5'
*.fal_server='r2'
*.log_archive_dest_1='LOCATION=/u01/archivelog  VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=r5'
*.log_archive_dest_2='SERVICE=r2 LGWR SYNC AFFIRM VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=r2'
*.log_archive_dest_3='SERVICE=r6 LGWR SYNC AFFIRM VALID_FOR=(STANDBY_LOGFILES,STANDBY_ROLE) DB_UNIQUE_NAME=r6'


ALTER SYSTEM SET LOG_ARCHIVE_DEST_3='SERVICE=r6 LGWR  ASYNC AFFIRM VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=r6' scope=both;
ALTER SYSTEM SET LOG_ARCHIVE_CONFIG='DG_CONFIG=(r2,r5,r6)'  scope=both;



开启时报：
SQL> startup nomount pfile='/home/oracle/initr2.ora'
ORA-01678: parameter db_file_name_convert must be pairs of pattern and replacement strings



Cascaded Redo Transport Destinations
A cascaded redo transport destination (also known as a terminal destination) receives primary database redo indirectly from a standby database rather than directly from a primary database.
A physical standby database that cascades primary database redo to one or more terminal destinations at the same time it is applying changes to its local database files is known as a cascading standby database.
With cascading, the overhead associated with performing redo transport is offloaded from a primary database to a cascading standby database.
A cascading standby database can cascade primary database redo to up to 30 terminal destinations.






官方文档安装如下：
Example 7–1 Some of the Initialization Parameters Used When Cascading Redo
Primary Database
DB_UNIQUE_NAME=boston
FAL_SERVER=boston2
LOG_ARCHIVE_CONFIG='DG_CONFIG=(boston,boston2,denver)'
LOG_ARCHIVE_DEST_1='LOCATION=USE_DB_RECOVERY_FILE_DEST VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=boston'
LOG_ARCHIVE_DEST_2='SERVICE=boston2 SYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) DB_UNIQUE_NAME=boston2'

Cascading Physical Standby Database
DB_UNIQUE_NAME=boston2
FAL_SERVER=boston
LOG_ARCHIVE_CONFIG= 'DG_CONFIG=(boston,boston2,denver)'
LOG_ARCHIVE_DEST_1='LOCATION= USE_DB_RECOVERY_FILE_DEST VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=boston2'
LOG_ARCHIVE_DEST_2= 'SERVICE=denver VALID_FOR=(STANDBY_LOGFILES,STANDBY_ROLE) DB_UNIQUE_NAME=denver'

Cascaded Physical Standby Database
DB_UNIQUE_NAME=denver
FAL_SERVER=boston2
LOG_ARCHIVE_CONFIG='DG_CONFIG=(boston,boston2,denver)'
LOG_ARCHIVE_DEST_1='LOCATION= USE_DB_RECOVERY_FILE_DEST VALID_FOR=(ALL_LOGFILES,ALL_ROLES) DB_UNIQUE_NAME=denver'

注意事项：

   1。备库r6需要重启f5才能同步数据，而且只能通过归档日志同步到r6

        r2 到 r5 是 归档日志与现线日志同步

        r5 到 r6 只能通过归档日志同步了，所以必须在r2切换日志，r6才会接受到日志并同步数据

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

  

  2。 官方文档说最好是11.2.0.2或以上的版本，经测试11.2.0.1也是可以的

如有错误麻烦提出来，谢谢！