mysql5.7 主从复制的正常切换【转】

目前环境如下：

master server IP:172.17.61.131

slave server IP:172.17.61.132

mysql version: mysql-5.7.21-linux

目标：计划内的主从复制正常切换

 

1.切换之前需要检查slave的同步状态，对master进行锁表。

slave server: 确保Slave_IO_Running和Slave_SQL_Running都为YES

[html] view plain copy

1. mysql> show slave status \G  
2. *************************** 1. row ***************************  
3.                Slave_IO_State: Waiting for master to send event  
4.                   Master_Host: 172.17.61.131  
5.                   Master_User: repl  
6.                   Master_Port: 3306  
7.                 Connect_Retry: 60  
8.               Master_Log_File: mysql_bin.000036  
9.           Read_Master_Log_Pos: 154  
10.                Relay_Log_File: slave_relay_bin.000003  
11.                 Relay_Log_Pos: 367  
12.         Relay_Master_Log_File: mysql_bin.000036  
13.              Slave_IO_Running: Yes  
14.             Slave_SQL_Running: Yes  
15.              ...  
16. 1 row in set (0.00 sec)  

master server进行锁表操作，由于我是测试环境没有应用，所以锁表很快。

2.在flush tables with read lock成功获得锁之前，必须等待所有语句执行完成（包括SELECT）。所以如果有个慢查询在执行，或者一个打开的事务，或者其他进程拿着表锁，flush tables with read lock就会被阻塞，直到所有的锁被释放。

[html] view plain copy

1. mysql> flush tables with read lock;  
2. Query OK, 0 rows affected (0.01 sec)  

3.接着要确保master所有的binlog已同步到slave，并且slave都已apply了所有的binlog。

maser server:

[html] view plain copy

1. mysql> show processlist;  
2. +----+------+---------------------+------+-------------+------+---------------------------------------------------------------+------------------+  
3. | Id | User | Host                | db   | Command     | Time | State                                                         | Info             |  
4. +----+------+---------------------+------+-------------+------+---------------------------------------------------------------+------------------+  
5. |  2 | repl | 172.17.61.132:60079 | NULL | Binlog Dump | 1907 | Master has sent all binlog to slave; waiting for more updates | NULL             |  
6. |  3 | root | localhost           | NULL | Query       |    0 | starting                                                      | show processlist |  
7. +----+------+---------------------+------+-------------+------+---------------------------------------------------------------+------------------+  
8. 2 rows in set (0.00 sec)  

slave server:

[html] view plain copy

1. mysql> show processlist;  
2. +----+-------------+-----------+------+---------+------+--------------------------------------------------------+------------------+  
3. | Id | User        | Host      | db   | Command | Time | State                                                  | Info             |  
4. +----+-------------+-----------+------+---------+------+--------------------------------------------------------+------------------+  
5. |  1 | system user |           | NULL | Connect | 1918 | Slave has read all relay log; waiting for more updates | NULL             |  
6. |  2 | system user |           | NULL | Connect | 1918 | Waiting for master to send event                       | NULL             |  
7. |  4 | root        | localhost | NULL | Query   |    0 | starting                                               | show processlist |  
8. +----+-------------+-----------+------+---------+------+--------------------------------------------------------+------------------+  
9. 3 rows in set (0.00 sec)  

4.停止slave的进程 

slave sever:

[html] view plain copy

1. mysql> STOP SLAVE IO_THREAD;  
2. Query OK, 0 rows affected (0.02 sec)  
3.   
4. mysql>  show processlist;  
5. +----+-------------+-----------+------+---------+------+--------------------------------------------------------+------------------+  
6. | Id | User        | Host      | db   | Command | Time | State                                                  | Info             |  
7. +----+-------------+-----------+------+---------+------+--------------------------------------------------------+------------------+  
8. |  1 | system user |           | NULL | Connect | 2039 | Slave has read all relay log; waiting for more updates | NULL             |  
9. |  4 | root        | localhost | NULL | Query   |    0 | starting                                               | show processlist |  
10. +----+-------------+-----------+------+---------+------+--------------------------------------------------------+------------------+  
11. 2 rows in set (0.00 sec)  

5.修改slave的my.cnf，重启mysql

[html] view plain copy

1. [mysqld]  
2. socket = /usr/local/mysql/mysql.sock  
3. character_set_server= utf8  
4. init_connect= 'SET NAMES utf8'  
5. basedir= /usr/local/mysql  
6. datadir= /u01/mysql  
7. socket = /u01/mysql/mysql.sock  
8. log-error= /u01/log/mysql/mysql_3306.err  
9. pid-file= /u01/mysql/mysqld.pid  
10. lower_case_table_names = 1  
11. sql_mode= STRICT_TRANS_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION  
12. secure-file-priv = /u01/backup  
13. server-id=10001  
14. log_bin = /u01/mysql/mysql_bin  
15. #skip-grant-tables  
16. innodb_flush_log_at_trx_commit=1  
17. sync_binlog=1  
18. #relay-log=/u01/mysql/slave_relay_bin  
19. expire_logs_days=10  
20. read_only=0  
21. #relay_log_recovery=on  
22. #relay_log_info_repository=TABLE  
23. max_binlog_size=1073741824  
24. #autocommit=off  
25. #long_query_time=15  
26. #slow_query_log=on  

[html] view plain copy

1. [root@qht132 ~]# service mysql restart  
2. Shutting down MySQL..                                      [  OK  ]  
3. Starting MySQL.                                            [  OK  ]  

需要开启原slave的bin_log,关闭relay_log*，关闭read_only

5.提升slave为master

[html] view plain copy

1. mysql> stop slave;  
2. Query OK, 0 rows affected (0.00 sec)  
3.   
4. mysql> reset master;  
5. Query OK, 0 rows affected (0.00 sec)  
6.   
7. mysql> reset slave all;  
8. Query OK, 0 rows affected (0.01 sec)  
9.   
10. mysql> show variables like 'read_only';  
11. +---------------+-------+  
12. | Variable_name | Value |  
13. +---------------+-------+  
14. | read_only     | OFF   |  
15. +---------------+-------+  
16. 1 row in set (0.00 sec)  
17.   
18. mysql> show master status \G  
19. *************************** 1. row ***************************  
20.              File: mysql_bin.000001  
21.          Position: 154  
22.      Binlog_Do_DB:  
23.  Binlog_Ignore_DB:  
24. Executed_Gtid_Set:  
25. 1 row in set (0.00 sec)  

在新的master上建立同步帐户并赋予同步权限（由于我为了省事，之前主从配置的时候将61网段的权限都赋予给了repl，所以这一步我也可以不做）

[html] view plain copy

1. mysql> grant replication slave on *.* to 'repl'@'172.17.61.%';  
2. Query OK, 0 rows affected (0.01 sec)  

6.修改原master的my.cnf，重启mysql

[html] view plain copy

1. [mysqld]  
2. socket = /usr/local/mysql/mysql.sock  
3. character_set_server= utf8  
4. init_connect= 'SET NAMES utf8'  
5. basedir= /usr/local/mysql  
6. datadir= /u01/mysql  
7. socket = /u01/mysql/mysql.sock  
8. log-error= /u01/log/mysql/mysql_3306.err  
9. pid-file= /u01/mysql/mysqld.pid  
10. lower_case_table_names = 1  
11. sql_mode= STRICT_TRANS_TABLES,NO_ZERO_IN_DATE,NO_ZERO_DATE,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION  
12. secure-file-priv = /u01/backup  
13. server-id=10000  
14. #log_bin = /u01/mysql/mysql_bin  
15. #skip-grant-tables  
16. #innodb_flush_log_at_trx_commit=1  
17. #sync_binlog=1  
18. expire_logs_days=10  
19. read_only=1  
20. relay_log_recovery=on  
21. relay_log_info_repository=TABLE  
22. #max_binlog_size=1073741824  
23. #autocommit=off  
24. #long_query_time=15  
25. #slow_query_log=on  

关闭原master的binlog，开启relay_log*，打开read_only状态

[html] view plain copy

1. [root@qht131 ~]# service mysql restart  
2. Shutting down MySQL..                                      [  OK  ]  
3. Starting MySQL.                                            [  OK  ]  

7.将master切换为slave

[html] view plain copy

1. mysql> reset master;  
2. Query OK, 0 rows affected (0.00 sec)  
3.   
4. mysql> change master to  
5.     -> master_host='172.17.61.132',  
6.     ->  master_user='repl',  
7.     ->  master_password='repl',  
8.     -> master_log_file='mysql_bin.000001',  
9.     -> master_log_pos=154;  
10. Query OK, 0 rows affected, 2 warnings (0.03 sec)  
11.   
12. mysql> start slave;  
13. Query OK, 0 rows affected (0.01 sec)  

master_log_file和master_log_pos需指定为新master切换后显示的起始位置

检查一下新slave的状态，确保slave_io_running和slave_sql_running都为YES。

[html] view plain copy

1. mysql> show slave status\G  
2. *************************** 1. row ***************************  
3.                Slave_IO_State: Waiting for master to send event  
4.                   Master_Host: 172.17.61.132  
5.                   Master_User: repl  
6.                   Master_Port: 3306  
7.                 Connect_Retry: 60  
8.               Master_Log_File: mysql_bin.000001  
9.           Read_Master_Log_Pos: 360  
10.                Relay_Log_File: qht131-relay-bin.000002  
11.                 Relay_Log_Pos: 526  
12.         Relay_Master_Log_File: mysql_bin.000001  
13.              Slave_IO_Running: Yes  
14.             Slave_SQL_Running: Yes  
15.               Replicate_Do_DB:  
16.           Replicate_Ignore_DB:  
17.            Replicate_Do_Table:  
18.        Replicate_Ignore_Table:  
19.       Replicate_Wild_Do_Table:  
20.   Replicate_Wild_Ignore_Table:  
21.                    Last_Errno: 0  
22.                    Last_Error:  
23.                  Skip_Counter: 0  
24.           Exec_Master_Log_Pos: 360  
25.               Relay_Log_Space: 734  
26.               Until_Condition: None  
27.                Until_Log_File:  
28.                 Until_Log_Pos: 0  
29.            Master_SSL_Allowed: No  
30.            Master_SSL_CA_File:  
31.            Master_SSL_CA_Path:  
32.               Master_SSL_Cert:  
33.             Master_SSL_Cipher:  
34.                Master_SSL_Key:  
35.         Seconds_Behind_Master: 0  
36. Master_SSL_Verify_Server_Cert: No  
37.                 Last_IO_Errno: 0  
38.                 Last_IO_Error:  
39.                Last_SQL_Errno: 0  
40.                Last_SQL_Error:  
41.   Replicate_Ignore_Server_Ids:  
42.              Master_Server_Id: 10001  
43.                   Master_UUID: 744cfcde-3a9b-11e8-b299-000c2900d025  
44.              Master_Info_File: /u01/mysql/master.info  
45.                     SQL_Delay: 0  
46.           SQL_Remaining_Delay: NULL  
47.       Slave_SQL_Running_State: Slave has read all relay log; waiting for more updates  
48.            Master_Retry_Count: 86400  
49.                   Master_Bind:  
50.       Last_IO_Error_Timestamp:  
51.      Last_SQL_Error_Timestamp:  
52.                Master_SSL_Crl:  
53.            Master_SSL_Crlpath:  
54.            Retrieved_Gtid_Set:  
55.             Executed_Gtid_Set:  
56.                 Auto_Position: 0  
57.          Replicate_Rewrite_DB:  
58.                  Channel_Name:  
59.            Master_TLS_Version:  
60. 1 row in set (0.00 sec)  

8.最后进行测试

新master:

[html] view plain copy

1. mysql> create table t3 ( c1 int);  
2. Query OK, 0 rows affected (0.05 sec)  

新slave:

[html] view plain copy

1. mysql> use l5m  
2. Database changed  
3. mysql> show tables;  
4. +---------------+  
5. | Tables_in_l5m |  
6. +---------------+  
7. | t1            |  
8. | t2            |  
9. | t3            |  
10. | test_emp      |  
11. | tt            |  
12. +---------------+  
13. 5 rows in set (0.00 sec)  
14.   
15. mysql> show create table t3\G  
16. *************************** 1. row ***************************  
17.        Table: t3  
18. Create Table: CREATE TABLE `t3` (  
19.   `c1` int(11) DEFAULT NULL  
20. ) ENGINE=InnoDB DEFAULT CHARSET=utf8  
21. 1 row in set (0.00 sec)  

至此测试完毕！

 

转自

mysql5.7 主从复制的正常切换 - CSDN博客

https://blog.csdn.net/jolly10/article/details/79877564