MySql NDB cluster replication配置

文章目录

• 单机部署：
• 1. 创建ndb_mgmd配置文件：
• 2. 创建ndbd配置文件
• 3. 创建mysqld配置文件
• 4. 安装初始数据库
• 5. 按顺序启动ndb_mgmd, ndbd, mysqld
• 6. 查看ndb cluster状态
• Geographic redundancy
• 分别修改A和B的mysqld配置文件并重启mysqld：
• 重启后进入mysql客户端查看master状态
• 分别重新配置A和B的slave
• 检查配置结果
• 通过在数据库中修改数据来检查配置是否生效

官方文档： https://dev.mysql.com/doc/refman/5.6/en/mysql-cluster.html

安装过程：略

单机部署：

1. 创建ndb_mgmd配置文件：

SC-1:/storage/ezohenz # cat m
[NDB_MGMD]
HostName=SC-1

[NDBD DEFAULT]
NoOfReplicas=2

[NDBD]
HostName=SC-1
Id=11
DataDir=/storage/ezohenz/data_ndbd1

[NDBD]
HostName=SC-1
Id=12
DataDir=/storage/ezohenz/data_ndbd2

[MYSQLD]
HostName=SC-1
Id=21

[MYSQLD]
HostName=SC-2
Id=22

2. 创建ndbd配置文件

SC-1:/storage/ezohenz # cat d
[MYSQLD]
ndbcluster
ndb-connectstring=SC-1:1186

[MYSQL_CLUSTER]
ndb-connectstring=SC-1:1186

3. 创建mysqld配置文件

SC-1:/storage/ezohenz # cat q

[MYSQLD]
ndbcluster
ndb-connectstring=SC-1:1186
default-storage-engine=ndbcluster
user=mysql

4. 安装初始数据库

mysql_install_db --user=mysql --basedir=/usr/local/mysql --datadir=./data_mysqld

5. 按顺序启动ndb_mgmd, ndbd, mysqld

ndb_mgmd -f ./m --initial
ndbd --defaults-file=./d --initial
mysqld --defaults-file=./q --datadir=./data_mysqld &

6. 查看ndb cluster状态

ndb_mgm -e show

注意貌似需要确认配置文件m中的所有ndbd都启动并连接上mgmd，mysqld才能正常连接上mgmd并创建engine为ndb的数据表。

Geographic redundancy

以上为单机cluster，需要多个才有实际意义。而在不同地区的cluster之间，也可以进行数据的同步。

现在假设有另一套单机cluster，配置如上面的1-6步。假设两个cluster分别为A和B，则：

分别修改A和B的mysqld配置文件并重启mysqld：

SC-1:/storage/db_sc1 # cat q

[MYSQLD]
ndbcluster
ndb-connectstring=SC-1:1186
default-storage-engine=ndbcluster
user=mysql

server-id=1
replicate-do-db=test
replicate-do-table=test.ctest
slave-skip-errors=1062,1032,1590
log-slave-updates
log-bin=
sync_binlog=1
binlog_format=MIXED
expire_logs_days=3
binlog-do-db=test
slave-net-timeout=10

务必将A和B的server-id配置成不同的值。

重启后进入mysql客户端查看master状态

执行show master status;

mysql> show master status;
+-----------------+----------+--------------+------------------+-------------------+
| File            | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+-----------------+----------+--------------+------------------+-------------------+
| SC-1-bin.000001 |      120 | test         |                  |                   |
+-----------------+----------+--------------+------------------+-------------------+
1 row in set (0.00 sec)

分别记录A和B的File，在下面的命令中会用到。

分别重新配置A和B的slave

在A (sc-1)的mysql客户端输入以下命令

stop slave;
change master to master_host='sc-2', master_log_file='SC-2-bin.000001', master_user='mysql', master_password='', master_retry_count=86400, master_connect_retry=5;
start slave;

在B (sc-2)的mysql客户端输入以下命令

stop slave;
change master to master_host='sc-1', master_log_file='SC-1-bin.000001', master_user='mysql', master_password='', master_retry_count=86400, master_connect_retry=5;
start slave;

此时将会在两个site的客户端上分别看到连接成功的消息。

检查配置结果

在A和B上分别执行show slave status\G，应该看到Slave_IO_Running和Slave_SQL_Running状态都是Yes：

show slave status\G
*************************** 1. row ***************************
               Slave_IO_State: Waiting for master to send event
                  Master_Host: sc-2
                  Master_User: mysql
                  Master_Port: 3306
                Connect_Retry: 5
              Master_Log_File: SC-2-bin.000001
          Read_Master_Log_Pos: 120
               Relay_Log_File: SC-1-relay-bin.000002
                Relay_Log_Pos: 282
        Relay_Master_Log_File: SC-2-bin.000001
             Slave_IO_Running: Yes
            Slave_SQL_Running: Yes
              Replicate_Do_DB: test
          Replicate_Ignore_DB:
           Replicate_Do_Table: test.ctest
       Replicate_Ignore_Table:
      Replicate_Wild_Do_Table:
  Replicate_Wild_Ignore_Table:
                   Last_Errno: 0
                   Last_Error:
                 Skip_Counter: 0
          Exec_Master_Log_Pos: 120
              Relay_Log_Space: 454
              Until_Condition: None
               Until_Log_File:
                Until_Log_Pos: 0
           Master_SSL_Allowed: No
           Master_SSL_CA_File:
           Master_SSL_CA_Path:
              Master_SSL_Cert:
            Master_SSL_Cipher:
               Master_SSL_Key:
        Seconds_Behind_Master: 0
Master_SSL_Verify_Server_Cert: No
                Last_IO_Errno: 0
                Last_IO_Error:
               Last_SQL_Errno: 0
               Last_SQL_Error:
  Replicate_Ignore_Server_Ids:
             Master_Server_Id: 2
                  Master_UUID: 5301a65f-34b9-11e8-a869-021020010200
             Master_Info_File: /cluster/storage/db_sc1/data_mysqld/master.info
                    SQL_Delay: 0
          SQL_Remaining_Delay: NULL
      Slave_SQL_Running_State: Slave has read all relay log; waiting for the slave I/O thread to update it
           Master_Retry_Count: 86400
                  Master_Bind:
      Last_IO_Error_Timestamp:
     Last_SQL_Error_Timestamp:
               Master_SSL_Crl:
           Master_SSL_Crlpath:
           Retrieved_Gtid_Set:
            Executed_Gtid_Set:
                Auto_Position: 0
1 row in set (0.00 sec)

通过在数据库中修改数据来检查配置是否生效

在A的数据库test中建一个ctest表，并在其中插入一条记录，应该能在B上看到插入的结果。反之亦然。

mysql> USE test;
Database changed
mysql> CREATE TABLE ctest (i INT) ENGINE=NDBCLUSTER;
2018-03-31 16:49:10 619 [Note] NDB Binlog: CREATE TABLE Event: REPL$test/ctest
2018-03-31 16:49:10 619 [Note] NDB Binlog: logging ./test/ctest (UPDATED,USE_WRITE)
Query OK, 0 rows affected (0.48 sec)

mysql> show tables;
+----------------+
| Tables_in_test |
+----------------+
| ctest          |
+----------------+
1 row in set (0.00 sec)

mysql> select * from ctest;
Empty set (0.00 sec)

mysql> insert into ctest values (111);
Query OK, 1 row affected (0.00 sec)

mysql> select * from ctest;
+------+
| i    |
+------+
|  111 |
+------+
1 row in set (0.00 sec)

全文完。