Dataguard for windows to linux（云上）异构加阿里云容灾配置

windows和linux平台可以实现跨平台的异构Dataguard，配置方法和同平台没有太大区别，注意目录的格式。

1、环境准备

1.1 主库开启归档

-----开启归档

alter system set LOG_ARCHIVE_DEST_1='LOCATION=/archivelog' scope=both;

shutdown immediate

start mount

alter database archivelog;

1.2 配置监听

--备库监听

SID_LIST_LISTENER =

  (SID_LIST =

    (SID_DESC =

     (GLOBAL_DBNAME = orcl)

     (ORACLE_HOME = /u01/app/oracle/product/11.2.0/db/)

     (SID_NAME = orcl_dr1)

    )

   )

主备库tns

ORCL_DR=

  (DESCRIPTION =

    (ADDRESS_LIST =

      (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.x.x)(PORT = 1521))

    )

    (CONNECT_DATA =

      (SERVICE_NAME = orcl_dr)

    )

  )

ORCL =

  (DESCRIPTION =

    (ADDRESS_LIST =

      (ADDRESS = (PROTOCOL = TCP)(HOST = 192.168.x.x)(PORT = 1521))

    )

    (CONNECT_DATA =

      (SERVICE_NAME = orcl)

    )

  )

1.3 密码文件

orapwd file=$ORACLE_HOME/dbs/orapworcl_dr1 password=oracle

2、Dupalicate 方式配置Dataguard

rman target sys/oracle@orcl auxiliary sys/oracle@orcl_dr

rman target sys/oracle@orcl auxiliary sys/oracle@orcl_dr

set newname for datafile 1 to '+DATA/orcl_dr/system01.dbf';

set newname for datafile 2 to '+DATA/orcl_dr/sysaux01.dbf';

set newname for datafile 3 to '+DATA/orcl_dr/undotbs01.dbf';

set newname for datafile 4 to '+DATA/orcl_dr/users01.dbf';

set newname for datafile 5 to '+FLASH/orcl_dr/testtbs01.dbf';

run{

set newname for tempfile 1 to '+DATA/orcl_dr/temp01.dbf';

set newname for datafile 1 to '+DATA/orcl_dr/system01.dbf';

set newname for datafile 2 to '+DATA/orcl_dr/sysaux01.dbf';

set newname for datafile 3 to '+DATA/orcl_dr/undotbs01.dbf';

set newname for datafile 4 to '+DATA/orcl_dr/users01.dbf';

set newname for datafile 5 to '+FLASH/orcl_dr/testtbs01.dbf';

DUPLICATE TARGET DATABASE

FOR STANDBY

FROM ACTIVE DATABASE

nofilenamecheck

dorecover

SPFILE

SET DB_UNIQUE_NAME="orcl_dr"

SET AUDIT_FILE_DEST="/u01/app/oracle/admin/orcl_dr/adump"

SET DIAGNOSTIC_DEST="/u01/app/oracle"

set log_archive_config='dg_config=(orcl,orcl_dr)'

set log_archive_dest_1='location=+FLASH '

SET LOG_ARCHIVE_DEST_2="service=orcl LGWR SYNC REGISTER VALID_FOR=(online_logfile,primary_role) "

SET FAL_SERVER="orcl"

SET FAL_CLIENT="orcl_dr"

SET CONTROL_FILES='+DATA/orcl_dr/control01.ctl','+DATA/orcl_dr/control02.ctl'

SET DB_FILE_NAME_CONVERT='C:\app\Administrator\oradata\orcl\','+DATA'

SET LOG_FILE_NAME_CONVERT='C:\app\Administrator\oradata\orcl\','+DATA'

SET DB_RECOVERY_FILE_DEST='+FLASH'

SET DB_RECOVERY_FILE_DEST_SIZE='1G';

}

部分输出：duplicate会输出它执行的命令和结果

---------- ''orcl_dr'' comment= '''' scope=spfile";    sql clone "alter system set  CONTROL_FILES = ''+DATA/orcl_dr/control01.ctl'', ''+DATA/orcl_dr/control02.ctl'' comment= '''' scope=spfile";    sql clone "alter system set  db_file_name_convert = ''C:\app\Administrator\oradata\orcl\'', ''+DATA'' comment= '''' scope=spfile";    sql clone "alter system set  LOG_FILE_NAME_CONVERT = ''C:\app\Administrator\oradata\orcl\'', ''+DATA'' comment= '''' scope=spfile";    sql clone "alter system set  db_recovery_file_dest = ''+FLASH'' comment= '''' scope=spfile";    sql clone "alter system set  DB_RECOVERY_FILE_DEST_SIZE = 1G comment= '''' scope=spfile";    shutdown clone immediate;    startup clone nomount; } 正在执行内存脚本 sql 语句: alter system set  db_unique_name =  ''orcl_dr'' comment= '''' scope=sp file sql 语句: alter system set  AUDIT_FILE_DEST =  ''/u01/app/oracle/admin/orcl_dr/a dump'' comment= '''' scope=spfile ………… 归档日志文件名=+FLASH/orcl_dr/archivelog/2017_01_19/thread_1_seq_38.261.93369654 9 线程=1 序列=38 介质恢复完成, 用时: 00:00:05 完成 recover 于 19-1月 -17 来自辅助数据库的 ORACLE 错误: ORA-01511: 重命名日志/数据文件时出错 ORA-01516: 不存在的日志文件, 数据文件或临时文件 "C:\APP\ADMINISTRATOR\ORADATA\OR CL\REDO01.LOG" RMAN-05535: 警告: 并非所有重做日志文件均已正确定义。 来自辅助数据库的 ORACLE 错误: ORA-01511: 重命名日志/数据文件时出错 ORA-01516: 不存在的日志文件, 数据文件或临时文件 "C:\APP\ADMINISTRATOR\ORADATA\OR CL\REDO02.LOG" RMAN-05535: 警告: 并非所有重做日志文件均已正确定义。 来自辅助数据库的 ORACLE 错误: ORA-01511: 重命名日志/数据文件时出错 ORA-01516: 不存在的日志文件, 数据文件或临时文件 "C:\APP\ADMINISTRATOR\ORADATA\OR CL\REDO03.LOG" RMAN-05535: 警告: 并非所有重做日志文件均已正确定义。 完成 Duplicate Db 于 19-1月 -17

复制过程中，可以通过sql命令查看通道进度

--rman进度查询

select sid,context,sofar,totalwork,round(sofar/totalwork*100,2) from v$session_longops where opname like 'RMAN%' and opname not like '%aggregate%' and totalwork<>0 and round(sofar/totalwork*100,2)<>100;

--复制过程中，主库做了添加数据文件的操作，以及复制暂停，只需要把脚本重新跑一下，根据复制的记录会重用已复制的数据文件

总结一下duplicate的过程：

1）连接目标库（主库）、辅助数据库（需要dbname即可）

2）进行参数配置

3）实例重启到mount

4）备份传输控制文件

5）复制数据文件：数据文件和临时数据文件分脚本进行

6）完成copy，编制文件目录

7）设置scn号，使用归档进行恢复

8）完成配置

3、主库参数配置

主库

alter system set standby_file_management=AUTO scope=both;

alter system set fal_server=orcl scope=both;

alter system set fal_client=orcl_dr scope=both;

alter system set LOG_ARCHIVE_DEST_2='SERVICE=orcl_dr LGWR SYNC VALID_FOR=(ONLINE_LOGFILES,PRIMARY_ROLE) NET_TIMEOUT=60 DB_UNIQUE_NAME=orcl_dr' scope=both;

alter system set log_archive_config='dg_config=(orcl,orcl_dr)' scope=both;

4、开启实时同步

--启用归档2

alter system set log_archive_dest_state_2='DEFER' scope=both  sid='*';

alter system set log_archive_dest_state_2='ENABLE' scope=both  sid='*';

alter database recover managed standby database disconnect from session;

--开启实时同步

ALTER DATABASE RECOVER MANAGED STANDBY DATABASE using current logfile disconnect from session;

取消同步

alter database recover managed standby database cancel;