CentOS 5.11下Oracle 11G R2 Dataguard搭建

Datagard算是Oracle企業(yè)版的一種容災(zāi)方案，在企業(yè)中廣泛應(yīng)用，我就將搭建過程記錄下來以作備用。

主機(jī)名    數(shù)據(jù)庫版本    實(shí)例名    IP

db1    Oracle 11G R2    member    172.16.1.250

db2    Oracle 11G R2    member    172.16.1.251

默認(rèn)情況下以上都已經(jīng)安裝好了Oracle數(shù)據(jù)庫，但是只在db1上建立了數(shù)據(jù)庫和監(jiān)聽，db2只安裝Oracle軟件不建庫，不建監(jiān)聽。

目錄：

1. 打開強(qiáng)制歸檔日志

2. 增加standby日志組

3. 修改主備啟動參數(shù)

4. 密碼文件的處理
   

5. 修改監(jiān)聽

6. 復(fù)制監(jiān)聽文件、參數(shù)文件、密碼文件到備庫
   

7. 創(chuàng)建備庫控制文件

8. 復(fù)制主庫數(shù)據(jù)文件和日志文件到備庫

9. 初始化及配置備庫做standby

10. DataGuard測試

11. 主備切換測試

基礎(chǔ)工作:

    a.安裝CentOS 5.11 x86_64,關(guān)閉selinux，iptables，自動對時(shí)

    b.安裝Oracle 11G R2,db1安裝軟件、監(jiān)聽及建庫,db2只安裝軟件不建庫 

    可以參考:http://fengwan.blog.51cto.com/508652/1330122

在db1的/etc/hosts里增加

127.0.0.1       db1

172.16.1.251    db2

在db2的/etc/hosts里增加

127.0.0.1       db2

172.16.1.250    db1

1. 打開強(qiáng)制歸檔(db1)

(db1)SQL >shutdown immediate;

(db1)SQL >startup mount;

(db1)SQL >alter database force logging; 

(db1)SQL >alter database archivelog;

2.創(chuàng)建重做日志組(必須要比原來的redo log多一組或多組,standby redo log是使用Real Time Apply的必要條件)

(db1)SQL> select group#,member from v$logfile;

    GROUP#    MEMBER

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

     3    /opt/oracle/oradata/member/redo03.log

     2    /opt/oracle/oradata/member/redo02.log

     1    /opt/oracle/oradata/member/redo01.log

從上面可以看出現(xiàn)在已經(jīng)有3組redo log.

(db1)SQL >alter database add standby logfile ('/opt/oracle/oradata/member/standby04.log') size 50m;

(db1)SQL >alter database add standby logfile ('/opt/oracle/oradata/member/standby05.log') size 50m;

(db1)SQL >alter database add standby logfile ('/opt/oracle/oradata/member/standby06.log') size 50m;

(db1)SQL >alter database add standby logfile ('/opt/oracle/oradata/member/standby07.log') size 50m;

再查下日志組是否創(chuàng)建成功

(db1)SQL> select group#,member from v$logfile;

    GROUP#    MEMBER

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

     3    /opt/oracle/oradata/member/redo03.log

     2    /opt/oracle/oradata/member/redo02.log

     1    /opt/oracle/oradata/member/redo01.log

    4    /opt/oracle/oradata/member/standby04.log

    5     /opt/oracle/oradata/member/standby05.log

    6     /opt/oracle/oradata/member/standby06.log

    7     /opt/oracle/oradata/member/standby07.log

可以看到我們創(chuàng)建的4組日志

3.修改主備庫的啟動參數(shù)

生成參數(shù)文件

(db1)SQL> create pfile='/tmp/member.pfile' from spfile;

退出sqlplus,用編輯器打開/tmp/member.pfile

[oracle@db1 ~]$ vi /tmp/member.pfile

member.__db_cache_size=331350016

member.__java_pool_size=4194304

member.__large_pool_size=4194304

member.__oracle_base='/opt/oracle'#ORACLE_BASE set from environment

member.__pga_aggregate_target=339738624

member.__sga_target=503316480

member.__shared_io_pool_size=0

member.__shared_pool_size=150994944

member.__streams_pool_size=0

*.audit_file_dest='/opt/oracle/admin/member/adump'

*.audit_trail='db'

*.compatible='11.2.0.0.0'

*.control_files='/opt/oracle/oradata/member/control01.ctl','/opt/oracle/flash_recovery_area/member/control02.ctl'

*.db_block_size=8192

*.db_domain=''

*.db_name='member'

*.db_recovery_file_dest='/opt/oracle/flash_recovery_area'

*.db_recovery_file_dest_size=4070572032

*.diagnostic_dest='/opt/oracle'

*.dispatchers='(PROTOCOL=TCP) (SERVICE=memberXDB)'

*.log_archive_format='%t_%s_%r.dbf'

*.memory_target=839909376

*.open_cursors=300

*.processes=150

*.remote_login_passwordfile='EXCLUSIVE'

*.undo_tablespace='UNDOTBS1'

#增加一下部分

*.db_unique_name='db1'

*.archive_lag_target=1800

*.fal_client='db1'

*.fal_server='db2'

*.log_archive_config='DG_CONFIG=(db1,db2)'

*.log_archive_dest_1='location=/opt/oracle/flash_recovery_area/ VALID_FOR=(all_logfiles,all_roles) db_unique_name=db1'

*.log_archive_dest_2='service=db2 lgwr async valid_for=(online_logfiles,primary_role) db_unique_name=db2'

*.log_archive_dest_state_1='enable'

*.log_archive_dest_state_2='enable'

*.log_archive_format='%t_%s_%r.dbf'

*.standby_file_management='auto'

*.db_file_name_convert='/opt/oracle/flash_recovery_area',' /opt/oracle/flash_recovery_area'

*.log_file_name_convert='/opt/oracle/flash_recovery_area ','/opt/oracle/flash_recovery_area'

并拷貝一個(gè)作為備機(jī)db2的啟動參數(shù)文件

[oracle@db1 ~]$ cp /tmp/member.pfile /tmp/db2.pfile

[oracle@db1 ~]$ vim /tmp/db2.pfile 

則將上面增加的部分修改為

*.db_unique_name='db2'

*.archive_lag_target=1800

*.fal_client='db2'

*.fal_server='db1'

*.log_archive_config='DG_CONFIG=(db1,db2)'

*.log_archive_dest_1='location=/opt/oracle/flash_recovery_area/ VALID_FOR=(all_logfiles,all_roles) db_unique_name=db2'

*.log_archive_dest_2='service=db1 lgwr async valid_for=(online_logfiles,primary_role) db_unique_name=db1'

*.log_archive_dest_state_1='enable'

*.log_archive_dest_state_2='enable'

*.log_archive_format='%t_%s_%r.dbf'

*.standby_file_management='auto'

*.db_file_name_convert='/opt/oracle/flash_recovery_area',' /opt/oracle/flash_recovery_area'

*.log_file_name_convert='/opt/oracle/flash_recovery_area ','/opt/oracle/flash_recovery_area'

在db1上以修改過的/tmp/member.pfile啟動

(db1)SQL >shutdown immediate;

(db1)SQL> startup pfile='/tmp/member.pfile' nomount;

(db1)SQL> create spfile from pfile='/tmp/member.pfile';

(db1)SQL >shutdown immediate;

(db1)SQL> startup;

4.主庫密碼文件:

[1]存在密碼文件

[oracle@db1 dbs]$ ls $ORACLE_HOME/dbs

hc_DBUA0.dat  hc_member.dat  init.ora  lkDB1  lkMEMBER  orapwmember  spfilemember.ora

看到上面有一個(gè)密碼文件orapwmember，在建庫的時(shí)候默認(rèn)會創(chuàng)建一個(gè)

=================================================================================

[2]不存在密碼文件

如果沒有的話可以手動創(chuàng)建一個(gè)。

[oracle@db1 dbs]$ cd $ORACLE_HOME/dbs

[oracle@db1 dbs]$ orapwd file=orapwmember password=123456 entries=3

#注意以上需要根據(jù)SID名建立的，file=orapwSID

===================================================================================

5.修改監(jiān)聽

[oracle@db1 dbs]$ cd $ORACLE_HOME/network/admin

[oracle@db1 admin]$ mv listener.ora listener.ora.default

[oracle@db1 admin]$ vim listener.ora

LISTENER =

  (DESCRIPTION =

    (ADDRESS = (PROTOCOL = TCP)(HOST = db1)(PORT = 1521))

  )

SID_LIST_LISTENER =

  (SID_LIST =

    (SID_DESC =

      (GLOBAL_DBNAME = member)

      (ORACLE_HOME = /opt/oracle/product/11.2.0/db_1)

      (SID_NAME = member)

    )

  )

#注意以上的HOST，就是HOSTNAME

[oracle@db1 admin]$ vim tnsnames.ora 

MEMBER =

  (DESCRIPTION =

    (ADDRESS = (PROTOCOL = TCP)(HOST = localhost.localdomain)(PORT = 1521))

    (CONNECT_DATA =

      (SERVER = DEDICATED)

      (SERVICE_NAME = member)

    )

  )

db1 =

  (DESCRIPTION =

    (ADDRESS_LIST =

      (ADDRESS = (PROTOCOL = TCP)(HOST = db1)(PORT = 1521))

    )

    (CONNECT_DATA =

      (SERVER = DEDICATED)

      (SERVICE_NAME = db1)

    )

  )

db2 =

  (DESCRIPTION =

    (ADDRESS_LIST =

      (ADDRESS = (PROTOCOL = TCP)(HOST = db2)(PORT = 1521))

    )

    (CONNECT_DATA =

      (SERVER = DEDICATED)

      (SERVICE_NAME = db2)

    )

  )

重啟下監(jiān)聽

[oracle@db1 admin]$ lsnrctl stop

[oracle@db1 admin]$ lsnrctl start

6.復(fù)制監(jiān)聽文件、參數(shù)文件、密碼文件到備庫

#注意一下我這邊$ORACLE_HOME

[oracle@db1 ~]$ echo $ORACLE_HOME

/opt/oracle/product/11.2.0/db_1

#傳輸啟動參數(shù)文件

[oracle@db1 ~]$ scp /tmp/db2.pfile db2:~

#傳輸密碼文件

[oracle@db1 ~]$ scp /opt/oracle/product/11.2.0/db_1/dbs/orapwmember db2:/opt/oracle/product/11.2.0/db_1/dbs/orapwmember

#傳輸監(jiān)聽文件

[oracle@db1 ~]$ scp -r /opt/oracle/product/11.2.0/db_1/network/admin/{listener.ora,tnsnames.ora} db2:/opt/oracle/product/11.2.0/db_1/network/admin/

在db2上修改/opt/oracle/product/11.2.0/db_1/network/admin/listener.ora,將db1修改為db2

[oracle@db2 ~]$ vim /opt/oracle/product/11.2.0/db_1/network/admin/listener.ora

LISTENER =

  (DESCRIPTION =

    (ADDRESS = (PROTOCOL = TCP)(HOST = db2)(PORT = 1521))

  ) 

SID_LIST_LISTENER =

  (SID_LIST =

    (SID_DESC =

      (GLOBAL_DBNAME = member)

      (ORACLE_HOME = /opt/oracle/product/11.2.0/db_1)

      (SID_NAME = member)

    )

  )

#只需要修改listener.ora即可，tnsnames.ora不需要動

7.主庫創(chuàng)建standby控制文件，我們這邊利用scp傳送全部文件

查看下控制文件的路徑

(db1)SQL> select name from v$controlfile;

NAME

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

/opt/oracle/oradata/member/control01.ctl

/opt/oracle/flash_recovery_area/member/control02.ctl

(db1)SQL> shutdown immediate;

(db1)SQL> startup mount;

(db1)SQL> alter database create standby controlfile as '/opt/oracle/oradata/member/standby.ctl';

#在/opt/oracle/oradata/member/目錄下創(chuàng)建standby.ctl備機(jī)控制文件

8.復(fù)制主庫數(shù)據(jù)文件和日志文件到備庫

[oracle@db1 ~]$ scp -r /opt/oracle/flash_recovery_area/ /opt/oracle/admin/ /opt/oracle/diag/ /opt/oracle/oradata/ db2:/opt/oracle

9.初始化備庫

在備機(jī)上使用standby的控制文件覆蓋原有的控制文件，覆蓋的路徑可以通過上一步查找控制文件的路徑了解到

[oracle@db2 ~]$ cp /opt/oracle/oradata/member/standby.ctl /opt/oracle/oradata/member/control01.ctl 

[oracle@db2 ~]$ cp /opt/oracle/oradata/member/standby.ctl /opt/oracle/flash_recovery_area/member/control02.ctl

使用db2.pfile之前修改過的參數(shù)文件進(jìn)行db2

(db2)SQL> startup pfile='/home/oracle/db2.pfile' nomount;

(db2)SQL> create spfile from pfile='/home/oracle/db2.pfile';

(db2)SQL> shutdown immediate;

(db2)SQL> startup nomount;

(db2)SQL> alter database mount standby database;

(db2)SQL> alter database open read only;

以下3種應(yīng)用日志的方法：（a和b選一）

a.開啟實(shí)時(shí)應(yīng)用日志，這樣在主庫插入立馬就可以在備機(jī)上查找到

(db2)SQL> alter database recover managed standby database using current logfile disconnect from session;

b.開啟redolog應(yīng)用日志，時(shí)間較長才能查詢到

(db2)SQL> alter database recover managed standby database disconnect from session;

c.停止應(yīng)用redolog，只接受日志，不重做

(db2)SQL> alter database recover managed standby database cancel;

至此，DataGuard搭建成功，在db1上創(chuàng)建表并插入數(shù)據(jù)，然后在db2上進(jìn)行查詢就可以查到了。一開始做的時(shí)候總是查不到數(shù)據(jù)，最后發(fā)現(xiàn)是上面應(yīng)用日志的方法問題。

10.Dataguard測試：

查看Standby管理進(jìn)程

(db1)SQL> select process,status from v$managed_standby;

PROCESS   STATUS

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

ARCH      CONNECTED

ARCH      CLOSING

ARCH      CLOSING

ARCH      CLOSING

LNS       WRITING

(db2)SQL> select process,status from v$managed_standby;

PROCESS   STATUS

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

ARCH      CLOSING

ARCH      CLOSING

ARCH      CONNECTED

ARCH      CONNECTED

MRP0      APPLYING_LOG

RFS       IDLE

RFS       IDLE

以上需要看到在主機(jī)上需要有LNS進(jìn)程，在備機(jī)上需要RFS進(jìn)程用來接收redo日志，MRP0進(jìn)程就是負(fù)責(zé)將日志寫入數(shù)據(jù)庫中

在db1上進(jìn)行切換日志，然后在db2上查看日志是否正常

(db1)SQL> select sequence#,applied from v$archived_log;

(db1)SQL> select max(sequence#) from v$archived_log;

MAX(SEQUENCE#)

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

    11

(db2)SQL> select sequence#,applied from v$archived_log;

(db2)SQL> select max(sequence#) from v$archived_log;

MAX(SEQUENCE#)

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

    11

(db1)SQL> alter system switch logfile;

(db1)SQL> select max(sequence#) from v$archived_log;

MAX(SEQUENCE#)

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

    12

(db2)SQL> select max(sequence#) from v$archived_log;

MAX(SEQUENCE#)

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

    12

從上面可以看出日志切換成功，DataGuard正常運(yùn)行

11.主備切換測試：

db1---primary/db2---standby   ===》db2--primary/db1--standby

[oracle@db1 ~]$ lsnrctl stop

(db1)SQL> alter database commit to switchover to physical standby with session shutdown;

(db1)SQL> shutdown immediate;

(db1)SQL> startup mount;

(db1)SQL> alter database open read only;

(db1)SQL> alter database recover managed standby database using current logfile disconnect from session;

//在執(zhí)行這條的時(shí)候，如果出現(xiàn)

ERROR at line 1:

ORA-01665: control file is not a standby control file

則是沒有執(zhí)行alter database commit to switchover to physical standby with session shutdown;

[oracle@db1 ~]$ lsnrctl start

(db2)SQL> alter database commit to switchover to primary;

(db2)SQL> shutdown immediate;

(db2)SQL> startup;

以上就是主備切換的流程