Oracle12c 容灾切换完全步骤

一. 切换前检查

1. 参数检查

----------------------------主端-----------------------------------------------------

--定义dataguard配置的有效db_unique_name

SQL> show parameter log_archive_config

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_config                   string      dg_config=(prod,prod_std)

-- 本地归档目的地及属性

SQL> show parameter log_archive_dest_1;

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_1                   string      location=+DATA_DG/arch valid_f
                                                 or=(all_logfiles,all_roles) db
                                                 _unique_name=prod

-- 远程归档目的地及属性

SQL> show parameter log_archive_dest_2

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_2                   string      service=prod_std lgwr async va
                                                 lid_for=(online_logfiles,prima
                                                 ry_roles) db_unique_name=prod_
                                                 std

--log_archive_dest_state_1

SQL> show parameter log_archive_dest_state_1

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_state_1             string      ENABLE

--log_archive_dest_state_2

SQL> show parameter log_archive_dest_state_2

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_state_2             string      ENABLE

--当主库数据文件被添加或删除，自动配置备库

SQL> show parameter standby_file_management

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
standby_file_management              string      auto

--备库TNS

SQL> show parameter fal_server

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
fal_server                           string      PROD_STD

--主库TNS

SQL> show parameter fal_client

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
fal_client                           string      PROD_PRI

--service_names

SQL> show parameter service_names;

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
service_names                        string      prod, prod_std

--spfile

SQL> show parameter spfile

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
spfile                               string      +DATA_DG/PROD/PARAMETERFILE/sp
                                                 file.281.950980793

----------------------------备端-----------------------------------------------------

--log_archive_config：与主端一致
--log_archive_dest_1

SQL> show parameter log_archive_dest_1

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_1                   string      location=+DATA_DG/arch valid_f
                                                 or=(all_logfiles,all_roles) db
                                                 _unique_name=prod_std

--log_archive_dest_2

SQL> show parameter log_archive_dest_2

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
log_archive_dest_2                   string      service=prod_pri lgwr async va
                                                 lid_for=(online_logfiles,prima
                                                 ry_roles) db_unique_name=prod

--log_archive_dest_state_1:与主端一致
--log_archive_dest_state_2:与主端一致
--standby_file_management:与主端一致
--fal_server

SQL> show parameter fal_server

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
fal_server                           string      PROD_PRI

--fal_client

SQL> show parameter fal_client

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
fal_client                           string      PROD_PRI

--db_unique_name

SQL> show parameter db_unique_name

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
db_unique_name                       string      prod_std

--spfile

SQL> show parameter spfile

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
spfile                               string      +DATA_DG/PROD/PARAMETERFILE/sp
                                                 fileprod.ora

2. 确认主备端网络连通情况

主端： 根据log_archive_dest_2的配置

[oracle@oracle-node1 ~]$ tnsping prod_std

TNS Ping Utility for Linux: Version 12.1.0.2.0 - Production on 03-AUG-2017 17:58:10

Copyright (c) 1997, 2014, Oracle.  All rights reserved.

Used parameter files:


Used TNSNAMES adapter to resolve the alias
Attempting to contact (DESCRIPTION_LIST = (LOAD_BALANCE = off) (FAILOVER = on) (DESCRIPTION = (ADDRESS_LIST = (LOAD_BALANCE = OFF) (FAILOVER = ON) (ADDRESS = (PROTOCOL = TCP)(HOST = 172.18.19.108)(PORT = 1521))) (CONNECT_DATA = (SERVICE_NAME = prod) (Instance_name = prod1) (FAILOVER_MODE = (TYPE = session) (METHOD = basic) (RETRIES = 4) (DELAY = 1)))) (DESCRIPTION = (ADDRESS_LIST = (LOAD_BALANCE = OFF) (FAILOVER = ON) (ADDRESS = (PROTOCOL = TCP)(HOST = 172.18.19.109)(PORT = 1521))) (CONNECT_DATA = (SERVICE_NAME = prod) (Instance_name = prod2) (FAILOVER_MODE = (TYPE = session) (METHOD = basic) (RETRIES = 4) (DELAY = 1)))))
OK (10 msec)

备端：

[oracle@oracle-node3 ~]$ tnsping prod_pri

TNS Ping Utility for Linux: Version 12.1.0.2.0 - Production on 03-AUG-2017 17:59:20

Copyright (c) 1997, 2014, Oracle.  All rights reserved.

Used parameter files:


Used TNSNAMES adapter to resolve the alias
Attempting to contact (DESCRIPTION_LIST = (LOAD_BALANCE = off) (FAILOVER = on) (DESCRIPTION = (ADDRESS_LIST = (LOAD_BALANCE = OFF) (FAILOVER = ON) (ADDRESS = (PROTOCOL = TCP)(HOST = 172.18.19.106)(PORT = 1521))) (CONNECT_DATA = (SERVICE_NAME = prod) (Instance_name = prod1) (FAILOVER_MODE = (TYPE = session) (METHOD = basic) (RETRIES = 4) (DELAY = 1)))) (DESCRIPTION = (ADDRESS_LIST = (LOAD_BALANCE = OFF) (FAILOVER = ON) (ADDRESS = (PROTOCOL = TCP)(HOST = 172.18.19.107)(PORT = 1521))) (CONNECT_DATA = (SERVICE_NAME = prod) (Instance_name = prod2) (FAILOVER_MODE = (TYPE = session) (METHOD = basic) (RETRIES = 4) (DELAY = 1)))))
OK (0 msec)

3. 主备端standby logfile查看,保证日志实时应用

SQL> select * from v$logfile where type='STANDBY';

4. 主备端数据库状态检查

---主端

SQL> select dbid,name,open_mode,current_scn,protection_mode,database_role,force_logging,
  2  open_mode,switchover_status from gv$database;

      DBID NAME      OPEN_MODE            CURRENT_SCN PROTECTION_MODE      DATABASE_ROLE    FORCE_LOGGING                 OPEN_MODE             SWITCHOVER_STATUS
---------- --------- -------------------- ----------- -------------------- ---------------- --------------------------------------- -------------------- --------------------
 362818713 PROD      READ WRITE               2262262 MAXIMUM PERFORMANCE  PRIMARY          YES                           READ WRITE           TO STANDBY
 362818713 PROD      READ WRITE               2262262 MAXIMUM PERFORMANCE  PRIMARY          YES                           READ WRITE            TO STANDBY

 --备端
 

SQL> select dbid,name,open_mode,current_scn,protection_mode,database_role,force_logging,
  2  open_mode,switchover_status from gv$database;

      DBID NAME      OPEN_MODE            CURRENT_SCN PROTECTION_MODE      DATABASE_ROLE    FORCE_LOGGING                 OPEN_MODE             SWITCHOVER_STATUS
---------- --------- -------------------- ----------- -------------------- ---------------- --------------------------------------- -------------------- --------------------
 362818713 PROD      READ ONLY WITH APPLY     2263439 MAXIMUM PERFORMANCE  PHYSICAL STANDBY YES                           READ ONLY WITH APPLYNOT ALLOWED
 362818713 PROD      READ ONLY WITH APPLY     2263439 MAXIMUM PERFORMANCE  PHYSICAL STANDBY YES                           READ ONLY WITH APPLY NOT ALLOWED

5. 检查容灾端应用节点和应用情况

SQL> select inst_id,value
  2    from gv$dataguard_stats
  3   where value =
  4         (select min(value) from gv$dataguard_stats where name = 'apply lag')
  5     and name = 'apply lag'
  6  ;

   INST_ID VALUE
---------- ----------------------------------------------------------------
         1 +00 00:00:00

二. switch over容灾切换

----------------------------主端-----------------------------------------------------

1. 关闭主端2号节点

--停掉两节点所有监听(关闭双节点所有监听，实时应用不缺少归档的情况下，生产的监听并不影响日志应用)

[root@oracle-node2 bin]# su - grid
Last login: Thu Aug  3 17:29:08 CST 2017 on pts/1
[grid@oracle-node2 ~]$ srvctl stop listener

--杀外部连接

[oracle@oracle-node2 ~]$ ps -ef|grep LOCAL=NO|awk '{print $2}'|xargs kill -9

--关闭数据库

[oracle@oracle-node2 ~]$ sqlplus / as sysdba
alter system switch logfile;
alter system switch logfile;
alter system switch logfile;
alter system switch logfile;
alter system switch logfile;
alter system switch logfile;
alter system switch logfile;
alter system switch logfile;
alter system checkpoint;
SQL> shutdown immediate;
Database closed.
Database dismounted.
ORACLE instance shut down.

2. 1号节点进行容灾切换

--杀外部连接

[root@oracle-node1 ~]# ps -ef|grep LOCAL=NO|awk '{print $2}'|xargs kill -9
--
SQL> alter system switch logfile;

System altered.

SQL> /

System altered.

SQL> /

System altered.

SQL> alter system checkpoint;

System altered.

SQL> alter database commit to switchover to physical standby with session shutdown;(执行主切备)

Database altered.
--
SQL> startup

--检查主库状态

SQL> select dbid,name,open_mode,current_scn,protection_mode,database_role,force_logging,
  2  open_mode,switchover_status from gv$database;

      DBID NAME      OPEN_MODE            CURRENT_SCN PROTECTION_MODE      DATABASE_ROLE    FORCE_LOGGING                           OPEN_MODE                 SWITCHOVER_STATUS
---------- --------- -------------------- ----------- -------------------- ---------------- --------------------------------------- -------------------- --------------------
 362818713 PROD      READ ONLY                2366359 MAXIMUM PERFORMANCE  PHYSICAL STANDBY YES
                                    READ ONLY                 TO PRIMARY

--开启日志应用

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

Database altered. 

3. 启动2号节点

SQL> startup

--检查主库状态，应为physical standby

SQL> select dbid,name,open_mode,current_scn,protection_mode,database_role,force_logging,
  2  open_mode,switchover_status from gv$database;

      DBID NAME      OPEN_MODE            CURRENT_SCN PROTECTION_MODE      DATABASE_ROLE    FORCE_LOGGING                           OPEN_MODE                 SWITCHOVER_STATUS
---------- --------- -------------------- ----------- -------------------- ---------------- --------------------------------------- -------------------- --------------------
 362818713 PROD      READ ONLY                2366359 MAXIMUM PERFORMANCE  PHYSICAL STANDBY YES                                     READ ONLY                 TO PRIMARY
 362818713 PROD      READ ONLY                2366359 MAXIMUM PERFORMANCE  PHYSICAL STANDBY YES                                     READ ONLY                 TO PRIMARY

--开启两节点监听

[grid@oracle-node2 ~]$ srvctl start listener

----------------------------备端-----------------------------------------------------

以日志应用在1号节点为例

1. 停容灾端2号节点

--停监听

[grid@oracle-node4 ~]$ srvctl stop listener -n oracle-node4

--杀外部连接

[root@oracle-node4 ~]# ps -ef|grep LOCAL=NO|awk '{print $2}'|xargs kill -9

--关闭数据库

SQL> shutdown abort;

2. 在容灾端1号节点进行备切主

--停监听

[grid@oracle-node3 ~]$ srvctl stop listener -n oracle-node3

--杀外部连接

[root@oracle-node3 ~]# ps -ef|grep LOCAL=NO|awk '{print $2}'|xargs kill -9

--检查数据库状态

SQL> select dbid,name,open_mode,current_scn,protection_mode,database_role,
  2  force_logging,switchover_status from v$database;

      DBID NAME      OPEN_MODE            CURRENT_SCN PROTECTION_MODE      DATABASE_ROLE    FORCE_LOGGING                    SWITCHOVER_STATUS
---------- --------- -------------------- ----------- -------------------- ---------------- --------------------------------------- --------------------
 362818713 PROD      READ ONLY WITH APPLY     2366359 MAXIMUM PERFORMANCE  PHYSICAL STANDBY YES                              TO PRIMARY

--取消日志应用

SQL> alter database recover managed standby database cancel;

Database altered.

--备转主，转换过程包括自动重启操作

SQL> alter database commit to switchover to primary with session shutdown;

Database altered.

--检查数据库状态，此时已经启动到mount

SQL> select dbid,name,open_mode,current_scn,protection_mode,database_role,
  2  force_logging,switchover_status from v$database;

      DBID NAME      OPEN_MODE            CURRENT_SCN PROTECTION_MODE      DATABASE_ROLE    FORCE_LOGGING                    SWITCHOVER_STATUS
---------- --------- -------------------- ----------- -------------------- ---------------- --------------------------------------- --------------------
 362818713 PROD      MOUNTED                        0 MAXIMUM PERFORMANCE  PRIMARY          YES                              NOT ALLOWED

3. 启动2号节点

--起库

SQL> startup

--开启监听

[grid@oracle-node4 ~]$ srvctl start listener

至此，switch over完成