MySQL集群是一个无共享的（share-nothing）、分布式节点架构的存储方案，目的是提供容错性和高性能。

MySQL NDB Cluster架构如图：

MySQL NDB Cluster架构按照节点类型分为三部分：

下面演示MySQL NDB Cluster的安装：

1、环境信息：

管理节点Mgmd：10.24.32.180
SQL节点Mysqld1：10.24.32.181
SQL节点Mysqld2：10.24.32.182
数据节点Ndbd1：10.24.32.183
数据节点Ndbd2：10.24.32.184

2、下载MySQL NDB Cluster软件；

[root@strong /]# ll /u01/
total 873104
-rw-r--r--. 1 root root 894037512 Mar  6 13:26 mysql-cluster-gpl-7.5.9-linux-glibc2.12-x86_64.tar.gz
[root@strong /]# 

3、安装SQL节点，在每个SQL节点都执行如下操作；

1）创建mysql用户和组，如果该用户和组不存在：

[root@strong ~]# grep mysql /etc/passwd
[root@strong ~]# grep mysql /etc/group
[root@strong ~]# groupadd mysql
[root@strong ~]# useradd -g mysql -s /bin/false mysql
[root@strong ~]# 

2）解压缩 mysql-cluster软件；

[root@strong local]# tar zxvf /u01/mysql-cluster-gpl-7.5.9-linux-glibc2.12-x86_64.tar.gz
[root@strong local]# ln -s mysql-cluster-gpl-7.5.9-linux-glibc2.12-x86_64/ mysql
[root@strong local]# 

3）设置数据库系统

[root@strong local]# cd mysql
[root@strong mysql]# ./bin/mysqld --initialize --user=mysql
2018-03-06T11:04:27.469235Z 0 [Warning] TIMESTAMP with implicit DEFAULT value is deprecated. Please use --explicit_defaults_for_timestamp server option (see documentation for more details).
2018-03-06T11:04:28.202258Z 0 [Warning] InnoDB: New log files created, LSN=45790
2018-03-06T11:04:28.291271Z 0 [Warning] InnoDB: Creating foreign key constraint system tables.
2018-03-06T11:04:28.360238Z 0 [Warning] No existing UUID has been found, so we assume that this is the first time that this server has been started. Generating a new UUID: 238f3d44-212e-11e8-a946-000c298da433.
2018-03-06T11:04:28.365955Z 0 [Warning] Gtid table is not ready to be used. Table 'mysql.gtid_executed' cannot be opened.
2018-03-06T11:04:28.371758Z 1 [Note] A temporary password is generated for root@localhost: 9gidWnDuEo_:

4）拷贝MySQL启动脚本；

[root@strong mysql]# cp support-files/mysql.server /etc/init.d/mysql.server

5）修改Root密码；

[root@strong mysql]# service mysql.server start
Starting MySQL.Logging to '/usr/local/mysql/data/strong.mysql.com.err'.
..... SUCCESS!
[root@strong mysql]# ./bin/mysql -uroot -p
Enter password:
Welcome to the MySQL monitor.  Commands end with ; or \g.
Your MySQL connection id is 5
Server version: 5.7.21-ndb-7.5.9-cluster-gplCopyright (c) 2000, 2018, Oracle and/or its affiliates. All rights reserved.Oracle is a registered trademark of Oracle Corporation and/or its
affiliates. Other names may be trademarks of their respective
owners.Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.mysql> alter user 'root'@'localhost' identified by 'root';
Query OK, 0 rows affected (0.00 sec)mysql> 

4、安装数据节点，在每个SQL节点都执行如下操作；

将ndbd、ndbmtd拷贝至/usr/local/bin/目录中：

[root@strong u01]# tar zxvf mysql-cluster-gpl-7.5.9-linux-glibc2.12-x86_64.tar.gz
[root@strong u01]# cp mysql-cluster-gpl-7.5.9-linux-glibc2.12-x86_64/bin/ndbd /usr/local/bin/ndbd
[root@strong u01]# cp mysql-cluster-gpl-7.5.9-linux-glibc2.12-x86_64/bin/ndbmtd /usr/local/bin/ndbmtd
[root@strong u01]# ll /usr/local/bin/
total 38280
-rwxr-xr-x. 1 root root 39197765 Mar  6 19:38 ndbd
-rwxr-xr-x. 1 root root        0 Mar  6 19:39 ndbmtd
[root@strong u01]# 

5、安装管理节点；

将ndb_mgm、ndb_mgmd拷贝至/usr/local/bin/目录中：

[root@strong local]# scp mysql-cluster-gpl-7.5.9-linux-glibc2.12-x86_64/bin/ndb_mgm* root@10.24.32.180:/usr/local/bin/
The authenticity of host '10.24.32.180 (10.24.32.180)' can't be established.
ECDSA key fingerprint is 5b:97:d1:f5:8b:84:a8:ac:9b:bf:b2:b4:0d:71:92:a5.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added '10.24.32.180' (ECDSA) to the list of known hosts.
root@10.24.32.180's password:
ndb_mgm                                                                                                             100% 7940KB   7.8MB/s   00:00
ndb_mgmd                                                                                                            100%   17MB  17.3MB/s   00:00
[root@strong local]# 

6、配置数据节点和SQL节点；

每个数据节点和SQL节点都需要一个my.cnf配置文件，该文件包括

配置信息：

[root@strong ~]# vi /etc/my.cnf
[mysqld]ndbcluster #run NDB storage engine[mysql_cluster]ndb-connectstring=10.24.32.180  #location of management server

7、配置管理节点；

创建目录，放置配置文件：

[root@strong local]# mkdir mysql
[root@strong local]# cd mysql/
[root@strong mysql]# mkdir mysql-cluster
[root@strong mysql-cluster]# vi config.ini
[ndbd default]#Opitions affecting ndbd processes on all nodes:NoOfReplicas=1  #Number of replicas[ndb_mgmd]HostName=10.24.32.180              #Hostname or IP address of MGM node
DataDir=/var/lib/mysql-cluster     #Dir for MGM node log files[ndbd]HostName=10.24.32.183
NodeId=2
DataDir=/u01[ndbd]HostName=10.24.32.184
NodeId=3
DataDir=/u01[mysqld]HostName=10.24.32.181[mysqld]HostName=10.24.32.182

8、启动Cluster；

Cluster需要各个节点都进行启动后才可以运行，节点的启动顺序为管理节点——>数据节点——>SQL节点。

启动管理节点：

[root@strong mysql-cluster]# ndb_mgmd -f ./config.ini
MySQL Cluster Management Server mysql-5.7.21 ndb-7.5.9
[root@strong mysql-cluster]# ps aux |grep ndb_mgmd
root      8155  2.5  0.6 504480  3380 ?        Ssl  20:48   0:00 ndb_mgmd -f ./config.ini
root      8166 13.0  0.1 112632   968 pts/0    S+   20:48   0:00 grep --color=auto ndb_mgmd
[root@strong mysql-cluster]# 

启动数据节点：

[root@strong local]# ndbd
2018-03-06 20:51:40 [ndbd] INFO     -- Angel connected to '10.24.32.180:1186'
2018-03-06 20:51:40 [ndbd] INFO     -- Angel allocated nodeid: 2

启动SQL节点：

[root@strong ~]# service mysql.server start
Starting MySQL........................................................................................................................................... SUCCESS! 

9、节点全部成功启动后，在管理节点查看集群状态；

[root@strong mysql-cluster]# ndb_mgm
-- NDB Cluster -- Management Client --
ndb_mgm> show
Connected to Management Server at: localhost:1186
Cluster Configuration
---------------------
[ndbd(NDB)] 2 node(s)
id=2   @10.24.32.183  (mysql-5.7.21 ndb-7.5.9, Nodegroup: 0, *)
id=3   @10.24.32.184  (mysql-5.7.21 ndb-7.5.9, Nodegroup: 1)[ndb_mgmd(MGM)]   1 node(s)
id=1   @10.24.32.180  (mysql-5.7.21 ndb-7.5.9)[mysqld(API)]   2 node(s)
id=4   @10.24.32.181  (mysql-5.7.21 ndb-7.5.9)
id=5   @10.24.32.182  (mysql-5.7.21 ndb-7.5.9)ndb_mgm> 

ndb_mgm是ndb_mgmd（MySQL Cluster Server）的客户端管理工具，通过它可以方便的检查Cluster的状态、启动备份、关闭Cluster等功能。

通过上面的状态信息，可以看到：

1）集群的管理服务端口是1186；
Connected to Management Server at: localhost:1186
2）集群的数据节点有2个；
[ndbd(NDB)] 2 node(s)
id=2   @10.24.32.183  (mysql-5.7.21 ndb-7.5.9, Nodegroup: 0, *)
id=3   @10.24.32.184  (mysql-5.7.21 ndb-7.5.9, Nodegroup: 1)
3）管理节点有一个；
[ndb_mgmd(MGM)] 1 node(s)
id=1   @10.24.32.180  (mysql-5.7.21 ndb-7.5.9)
4）SQL节点有2个，且处于连接状态；
[mysqld(API)]   2 node(s)
id=4   @10.24.32.181  (mysql-5.7.21 ndb-7.5.9)
id=5   @10.24.32.182  (mysql-5.7.21 ndb-7.5.9)

10、测试Cluster；

使用Cluster，则表的存储引擎必须为NDB，其他类型存储引擎的数据不会保存到数据节点中，对于CLuster的一个重要功能就是防止单点故障，下面将分别进行测试：

1）NDB存储引擎测试

--在任一SQL节点（本例用10.24.32.181）的test库创建测试表t_cluster，设置存储引擎为NDB：
mysql> create table t_cluster(id int,name varchar(100),cdate datetime default now()) engine=NDB;
Query OK, 0 rows affected (0.42 sec)mysql> insert into t_cluster(id,name) values(1,'Alen'),(2,'Cluster Test');
Query OK, 2 rows affected (0.10 sec)
Records: 2  Duplicates: 0  Warnings: 0mysql> select *from t_cluster;
+------+--------------+---------------------+
| id   | name         | cdate               |
+------+--------------+---------------------+
|    1 | Alen         | 2018-03-06 23:07:50 |
|    2 | Cluster Test | 2018-03-06 23:07:50 |
+------+--------------+---------------------+
2 rows in set (0.03 sec)mysql>
--在另一SQL节点（本例用10.24.32.182）的test库查询表t_cluster，结果如下：
mysql> select *from t_cluster;
+------+--------------+---------------------+
| id   | name         | cdate               |
+------+--------------+---------------------+
|    1 | Alen         | 2018-03-06 23:07:50 |
|    2 | Cluster Test | 2018-03-06 23:07:50 |
+------+--------------+---------------------+
2 rows in set (0.02 sec)mysql>
--如上，两个SQL节点查询到的数据是一致的。
--在SQL节点10.24.32.181将表t_cluster的存储引擎改为MyISAM，再次插入测试：
mysql> alter table t_cluster engine=myisam;
Query OK, 2 rows affected (0.66 sec)
Records: 2  Duplicates: 0  Warnings: 0mysql> insert into t_cluster(id,name) values(3,'MyISAM Test');
Query OK, 1 row affected (0.03 sec)mysql> select *from t_cluster;
+------+--------------+---------------------+
| id   | name         | cdate               |
+------+--------------+---------------------+
|    2 | Cluster Test | 2018-03-06 23:07:50 |
|    1 | Alen         | 2018-03-06 23:07:50 |
|    3 | MyISAM Test  | 2018-03-06 23:15:02 |
+------+--------------+---------------------+
3 rows in set (0.00 sec)mysql>
--在SQL节点10.24.32.182再次查询表t_cluster，可以发现，表无法查询，结果如下：
mysql> select *from t_cluster;
ERROR 1412 (HY000): Table definition has changed, please retry transaction
mysql>
--在SQL节点10.24.32.181将表t_cluster的存储引擎改为NDB：
mysql> alter table t_cluster engine=ndb;
Query OK, 3 rows affected (0.32 sec)
Records: 3  Duplicates: 0  Warnings: 0mysql>
--在SQL节点10.24.32.182再次查询，结果如下：
mysql> select *from t_cluster;
+------+--------------+---------------------+
| id   | name         | cdate               |
+------+--------------+---------------------+
|    3 | MyISAM Test  | 2018-03-06 23:15:02 |
|    1 | Alen         | 2018-03-06 23:07:50 |
|    2 | Cluster Test | 2018-03-06 23:07:50 |
+------+--------------+---------------------+
3 rows in set (0.01 sec)mysql>
--显然，表t_cluster的数据再次同步到了数据节点，所有SQL节点又都可以正常查询数据。

2）单点故障测试

对于任一种节点，都存在单点故障的可能性，在Cluster的设置过程中，应该尽量对每一类节点设置冗余，以防止单点故障发生时造成的应用中断；对于管理节点，一般不需要特殊的设置，只需要将管理工具和配置文件放在多台主机上即可。

SQL节点发生单点故障

--将SQL节点10.24.32.181的MySQL服务停止：
[root@strong mysql]# mysqladmin -uroot -p shutdown
Enter password:
--查看CLuster的状态，可以发现SQL节点10.24.32.181已经断开，另一SQL节点10.24.32.182仍然正常连接：
[root@strong mysql-cluster]# ndb_mgm
-- NDB Cluster -- Management Client --
ndb_mgm> show
Connected to Management Server at: localhost:1186
Cluster Configuration
---------------------
[ndbd(NDB)] 2 node(s)
id=2   @10.24.32.183  (mysql-5.7.21 ndb-7.5.9, Nodegroup: 0, *)
id=3   @10.24.32.184  (mysql-5.7.21 ndb-7.5.9, Nodegroup: 1)[ndb_mgmd(MGM)]   1 node(s)
id=1   @10.24.32.180  (mysql-5.7.21 ndb-7.5.9)[mysqld(API)]   2 node(s)
id=4 (not connected, accepting connect from 10.24.32.181)
id=5   @10.24.32.182  (mysql-5.7.21 ndb-7.5.9)ndb_mgm>
--从节点10.24.32.182查看t_cluster表，结果如下，显然，SQL节点的单点故障并没有引起数据的查询故障，对于应用来说，需要改变的就是将以前对故障节点的访问改为对非故障节点的访问：
mysql> select *from t_cluster;
+------+--------------+---------------------+
| id   | name         | cdate               |
+------+--------------+---------------------+
|    2 | Cluster Test | 2018-03-06 23:07:50 |
|    3 | MyISAM Test  | 2018-03-06 23:15:02 |
|    1 | Alen         | 2018-03-06 23:07:50 |
+------+--------------+---------------------+
3 rows in set (0.00 sec)

数据节点的单点故障

在该测试环境中，数据节点有两个，那么它们对数据的存储是互相镜像还是一份数据分成几块呢？这个答案的关键在于配置文件中[ndbd default]组中的noofreplicas参数，如果该值为1，表示只有一份数据，但是分成n块分别存储在n个数据节点上；如果等于2，则表示数据被分成n/2块，每块数据都有两个备份，这样即使有任一个节点发生故障，只要它的备份节点正常，系统就可以正常运行。

--管理节点10.24.32.180的[ndbd default]组中的noofreplicas参数为1，将数据节点10.24.32.183的NDB进程停止：
[root@strong u01]# ps -ef |grep ndbd
root      8253     1  0 20:51 ?        00:00:01 ndbd
root      8254  8253  3 20:51 ?        00:03:36 ndbd
root      8515  8336  0 22:48 pts/1    00:00:00 grep --color=auto ndbd
[root@strong u01]# kill 8254 8253
--查看集群状态：
ndb_mgm> show
Cluster Configuration
---------------------
[ndbd(NDB)] 2 node(s)
id=2 (not connected, accepting connect from 10.24.32.183)
id=3 (not connected, accepting connect from 10.24.32.184)[ndb_mgmd(MGM)]   1 node(s)
id=1   @10.24.32.180  (mysql-5.7.21 ndb-7.5.9)[mysqld(API)]   2 node(s)
id=4 (not connected, accepting connect from 10.24.32.181)
id=5 (not connected, accepting connect from 10.24.32.182)
--在任一SQL节点（10.24.32.181）查询表t_cluster，结果如下：
mysql> select *from t_cluster;
ERROR 1296 (HY000): Got error 4009 'Cluster Failure' from NDBCLUSTER
mysql>
--将noofreplicas参数改为2，按照前面的步骤重启集群：
NoOfReplicas=2
--停止数据节点10.24.32.183的NDB进程：
[root@strong ~]# ps aux |grep ndbd
root      1322  0.0  0.0  28584   128 ?        Ss   01:22   0:00 ndbd --initial
root      1323  1.8 56.6 1136400 281004 ?      Sl   01:22   0:04 ndbd --initial
root      1382  0.0  0.1 112632   964 pts/0    S+   01:26   0:00 grep --color=auto ndbd
[root@strong ~]# kill 1323
[root@strong ~]# ps aux |grep ndbd
root      1386  0.0  0.1 112632   964 pts/0    R+   01:26   0:00 grep --color=auto ndbd
[root@strong ~]#
--查看集群状态：
ndb_mgm> show
Cluster Configuration
---------------------
[ndbd(NDB)] 2 node(s)
id=2 (not connected, accepting connect from 10.24.32.183)
id=3   @10.24.32.184  (mysql-5.7.21 ndb-7.5.9, Nodegroup: 0, *)[ndb_mgmd(MGM)]    1 node(s)
id=1   @10.24.32.180  (mysql-5.7.21 ndb-7.5.9)[mysqld(API)]   2 node(s)
id=4   @10.24.32.181  (mysql-5.7.21 ndb-7.5.9)
id=5   @10.24.32.182  (mysql-5.7.21 ndb-7.5.9)ndb_mgm>
--再次从任一SQL节点10.24.32.181查询，结果如下：
mysql> select *from t_cluster;
+------+--------------+---------------------+
| id   | name         | cdate               |
+------+--------------+---------------------+
|    1 | Alen         | 2018-03-07 01:40:04 |
|    3 | MyISAM Test  | 2018-03-07 01:40:19 |
|    2 | Cluster Test | 2018-03-07 01:40:04 |
+------+--------------+---------------------+
3 rows in set (0.01 sec)

11、Cluster的关闭；

[root@strong mysql-cluster]# ndb_mgm
-- NDB Cluster -- Management Client --
ndb_mgm> shutdown
Connected to Management Server at: localhost:1186
Node 2: Cluster shutdown initiated
Node 3: Cluster shutdown initiated
Node 3: Node shutdown completed.
3 NDB Cluster node(s) have shutdown.
Disconnecting to allow management server to shutdown.
Node 2: Node shutdown completed.
ndb_mgm> 

注意：集群关闭后，SQL节点不会关闭，需手工关闭；