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

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。

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:

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:

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:

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

5.修改slave的my.cnf，重启mysql

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

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

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

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

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

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

7.将master切换为slave

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

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:

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

新slave:

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. mysql> show create table t3\G
15. *************************** 1. row ***************************
16. Table: t3
17. Create Table: CREATE TABLE `t3` (
18. `c1` int(11) DEFAULT NULL
19. ) ENGINE=InnoDB DEFAULT CHARSET=utf8
20. 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