概述

本章，您将学习到如何在一个运行的 MySQL NDB Cluster 中添加新的 SQL 节点。

在添加新 SQL 节点之前，先取消文件系统加密。

Shell (192.168.100.10)> killall ndb_mgmd

Shell (192.168.100.10)> vim /etc/mysql-cluster/config.ini
[ndbd default]
NoOfReplicas=2
DataMemory=98M
CompressedBackup=1
BackupDataDir=/mysql-cluster-backup
...

Shell (192.168.100.10)> ndb_mgmd --configdir=/etc/mysql-cluster/ -f /etc/mysql-cluster/config.ini --initial

# 节点 ID 为 2 的终端
Shell (192.168.100.12)> killall ndbd
Shell (192.168.100.12)> vim /etc/my.cnf
[mysqld]
ndbcluster

[mysql_cluster]
ndb-connectstring=192.168.100.10

Shell (192.168.100.12)> ndbd --initial

# 节点 ID 为 3 的终端
Shell (192.168.100.14)> killall ndbd
Shell (192.168.100.14)> vim /etc/my.cnf
[mysqld]
ndbcluster

[mysql_cluster]
ndb-connectstring=192.168.100.10

Shell (192.168.100.14)> ndbd --initial

# 节点 ID 为 4 的终端
Shell (192.168.100.18)> killall ndbd
Shell (192.168.100.18)> vim /etc/my.cnf
[mysqld]
ndbcluster

[mysql_cluster]
ndb-connectstring=192.168.100.10

Shell (192.168.100.18)> ndbd --initial

# 节点 ID 为 5 的数据数据
Shell (192.168.100.20)> killall ndbd
Shell (192.168.100.20)> vim /etc/my.cnf
[mysqld]
ndbcluster

[mysql_cluster]
ndb-connectstring=192.168.100.10

Shell (192.168.100.20)> ndbd --initial

# 重启 SQL 节点
Shell (192.168.100.16)> /usr/local/mysql/support-files/mysql.server restart

操作步骤

配置新 SQL 节点

依然是熟悉的操作：

# 添加无任何用户成员的组
Shell (192.168.100.22)> groupadd -r mysql

# 添加用户，并指定一些用户的属性
Shell (192.168.100.22)> useradd -r -g mysql -s /sbin/nologin mysql

# 下载
Shell (192.168.100.22)> wget https://dev.mysql.com/get/Downloads/MySQL-Cluster-8.4/mysql-cluster-8.4.8-linux-glibc2.28-x86_64.tar.xz

# 解压归档文件，并从归档文件中提取出文件
Shell (192.168.100.22)> tar -vxf mysql-cluster-8.4.8-linux-glibc2.28-x86_64.tar.xz -C /usr/local/src/

# 创建目录的软链接（符号链接）
Shell (192.168.100.22)> ln -s /usr/local/src/mysql-cluster-8.4.8-linux-glibc2.28-x86_64/ /usr/local/mysql

# 查阅
Shell (192.168.100.22)> ls -l /usr/local/mysql
lrwxrwxrwx 1 root root 58 Apr 26 14:50 /usr/local/mysql -> /usr/local/src/mysql-cluster-8.4.8-linux-glibc2.28-x86_64/

# 初始化，这将随机生成 root 的用户密码
## 若提示错误，请安装 libaio 软件包
Shell > cd /usr/local/mysql/bin/ && ./mysqld --initialize
2026-04-26T06:51:23.543661Z 0 [System] [MY-015017] [Server] MySQL Server Initialization - start.
2026-04-26T06:51:23.545008Z 0 [System] [MY-013169] [Server] /usr/local/src/mysql-cluster-8.4.8-linux-glibc2.28-x86_64/bin/mysqld (mysqld 8.4.8-cluster) initializing of server in progress as process 41207
2026-04-26T06:51:23.550415Z 1 [System] [MY-013576] [InnoDB] InnoDB initialization has started.
2026-04-26T06:51:23.800269Z 1 [System] [MY-013577] [InnoDB] InnoDB initialization has ended.
2026-04-26T06:51:25.694169Z 6 [Note] [MY-010454] [Server] A temporary password is generated for root@localhost: xvNa)b3=:xyf
2026-04-26T06:51:27.511061Z 0 [System] [MY-015018] [Server] MySQL Server Initialization - end.

# 递归变更链接目录下所有文件的所有者和所属组，递归变更数据目录的所有者
Shell (192.168.100.22)> chown -R root:mysql /usr/local/mysql/
Shell (192.168.100.22)> chown -R mysql /usr/local/mysql/data/

# 编辑 /etc/my.cnf 文件
Shell (192.168.100.22)> vim /etc/my.cnf
[mysqld]
ndbcluster
default-storage-engine=NDBCLUSTER

[mysql_cluster]
ndb-connectstring=192.168.100.10

对管理节点进行配置并执行滚动重启

在管理节点 config.ini 中配置相关参数。若有多个管理服务器（管理节点），则需要修改所有的 config.ini 文件：

Shell (192.168.100.10)> vim /etc/mysql-cluster/config.ini
[ndbd default]
NoOfReplicas=2
DataMemory=98M
CompressedBackup=1
BackupDataDir=/mysql-cluster-backup

[ndb_mgmd]
HostName=192.168.100.10
DataDir=/var/log/mysql-cluster
NodeId=1

[ndbd]
HostName=192.168.100.12
NodeId=2
DataDir=/usr/local/mysql/data

[ndbd]
HostName=192.168.100.14
NodeId=3
DataDir=/usr/local/mysql/data

[ndbd]
HostName=192.168.100.18
NodeId=4
DataDir=/usr/local/mysql/data

[ndbd]
HostName=192.168.100.20
NodeId=5
DataDir=/usr/local/mysql/data

[mysqld]
HostName=192.168.100.16
NodeId=6

[mysqld]
HostName=192.168.100.22
NodeId=7

[mysqld]

[mysqld]

滚动重启所有的管理服务器（管理节点），单个管理服务器（管理节点）的操作如下所示：

Shell (192.168.100.10)> killall ndb_mgmd

Shell (192.168.100.10)> ndb_mgmd --configdir=/etc/mysql-cluster/ -f /etc/mysql-cluster/config.ini --initial

注意！实际生产环境下，若您有多个管理节点，则不应该一次性将所有管理节点关闭，而应该是滚动重启一个管理节点后，再配置下一个管理节点。

对每一个数据节点执行滚动重启

# 在管理节点的终端中使用交互命令完成数据节点的滚动重启
Shell (192.168.100.10)> ndb_mgm
ndb_mgm> 2 restart
ndb_mgm> 3 restart
ndb_mgm> 4 restart
ndb_mgm> 5 restart

对每一个旧的SQL节点执行重启

Shell (192.168.100.16)> /usr/local/mysql/support-files/mysql.server restart

启动新SQL节点的进程并修改密码

Shell (192.168.100.22)> /usr/local/mysql/support-files/mysql.server start

Shell (192.168.100.22)> /usr/local/mysql/bin/mysql -h localhost -u root --password="xvNa)b3=:xyf"

MySQL > alter user 'root'@'localhost' identified by 'PassOut,15:0->';

最后

在任意一个管理节点上查看并验证：

Shell (192.168.100.10)> ndb_mgm -e "show"
[ndbd(NDB)]     4 node(s)
id=2    @192.168.100.12  (mysql-8.4.8 ndb-8.4.8, Nodegroup: 0, *)
id=3    @192.168.100.14  (mysql-8.4.8 ndb-8.4.8, Nodegroup: 0)
id=4    @192.168.100.18  (mysql-8.4.8 ndb-8.4.8, Nodegroup: 1)
id=5    @192.168.100.20  (mysql-8.4.8 ndb-8.4.8, Nodegroup: 1)

[ndb_mgmd(MGM)] 1 node(s)
id=1    @192.168.100.10  (mysql-8.4.8 ndb-8.4.8)

[mysqld(API)]   4 node(s)
id=6    @192.168.100.16  (mysql-8.4.8 ndb-8.4.8)
id=7    @192.168.100.22  (mysql-8.4.8 ndb-8.4.8)
id=8 (not connected, accepting connect from any host)
id=9 (not connected, accepting connect from any host)

赞赏