2021-06-20

搭建 MySQL 高可用高性能集群

什么是MySQL集群，什么是MySQL集群，如果你想知道什么是MySQL集群，我现在就带你研究。

MySQL 是一款流行的轻量级数据库，很多应用都是使用它作为数据存储。作为小型应用的数据库，它完全可以胜任，但是如果是大型应用，高性能高可用的要求，单服务器部署的MySQL就不够了。MySQL NDB Cluster 为这个需求提供了一个官方的集群解决方案。

MySQL NDB Cluster 是什么

MySQL NDB Cluster 是 MySQL 的一个高可用、高冗余版本，适用于分布式计算环境。
文档链接

搭建集群的前置工作

至少准备 3 台服务器，一台作为管理服务器，两台作为数据服务器和 SQL 服务器，当然有更多的服务器会更好。

管理服务器mgm：192.168.0.105数据服务器ndb1：192.168.0.106数据服务器ndb2：192.168.0.104sql服务器：192.168.0.106sql服务器：192.168.0.104

本文以 ubuntu20.04 为例，所有操作都可用于 ubuntu 系统。

开始部署集群

首先下载 MySQL NDB Cluster二进制文件，解压缩后开始下面的步骤。

部署管理服务器

更新系统

apt update -y && apt upgrade -y && apt install libncurses5 -y

复制 ndb_mgm 和 ndb_mgmd 到管理服务器

scp ./mysql-cluster-gpl-7.6.17-linux-glibc2.12-x86_64/bin/ndb_mgm* mgm@192.168.0.105:/home/mgm

在管理服务器复制 ndb_mgm 和 ndb_mgmd 到/usr/local/bin 文件夹

cp -rfv /home/mgm/ndb_mgm* /usr/local/bin

赋予 ndb_mgm 和 ndb_mgmd 可执行权限

chmod +x /usr/local/bin/ndb_mgm*

添加配置文件

mkdir /var/lib/mysql-clustervi /var/lib/mysql-cluster/config.ini

config.ini

[ndbd default]# Options affecting ndbd processes on all data nodes:NoOfReplicas=2      # Number of fragment replicasDataMemory=98M      # How much memory to allocate for data storage[ndb_mgmd]# Management process options:HostName=192.168.0.105    # Hostname or IP address of management nodeNodeId=1       # Node ID for this Management nodeDataDir=/var/lib/mysql-cluster  # Directory for management node log files[ndbd]# Options for data node "A":         # (one [ndbd] section per data node)HostName=192.168.0.104   # Hostname or IP addressNodeId=2       # Node ID for this data nodeDataDir=/data/mysql-cluster/data   # Directory for this data node's data files[ndbd]# Options for data node "B":        # (one [ndbd] section per data node)HostName=192.168.0.106   # Hostname or IP addressNodeId=3      # Node ID for this data nodeDataDir=/data/mysql-cluster/data  # Directory for this data node's data files[mysqld]# SQL node options:HostName=192.168.0.104  # Hostname or IP address        # (additional mysqld connections can be        # specified for this node for various        # purposes such as running ndb_restore)[mysqld]# SQL node options:HostName=192.168.0.106  # Hostname or IP address        # (additional mysqld connections can be        # specified for this node for various        # purposes such as running ndb_restore)

开启防火墙，集群管理服务默认使用 1186 端口

ufw allow 22ufw allow 1186ufw enable

初始化并启动管理服务器

cd /usr/local/bin/ndb_mgmd --initial --configdir=/var/lib/mysql-cluster -f /var/lib/mysql-cluster/config.ini --ndb-nodeid=1

当出现以下结果的时候，表示管理服务器已经启动成功了

root@mgm:/usr/local/bin# ndb_mgmd --initial --configdir=/var/lib/mysql-cluster -f /var/lib/mysql-cluster/config.ini --ndb-nodeid=1MySQL Cluster Management Server mysql-5.7.33 ndb-7.6.17

我们再执行 ndb_mgm 命令，可以查看当前集群的状态

root@mgm:/usr/local/bin# ndb_mgm-- NDB Cluster -- Management Client --ndb_mgm> showConnected to Management Server at: localhost:1186Cluster Configuration---------------------[ndbd(NDB)]	2 node(s)id=2 (not connected, accepting connect from 192.168.0.104)id=3 (not connected, accepting connect from 192.168.0.106)[ndb_mgmd(MGM)]	1 node(s)id=1	@192.168.0.105 (mysql-5.7.33 ndb-7.6.17)[mysqld(API)]	2 node(s)id=4 (not connected, accepting connect from 192.168.0.104)id=5 (not connected, accepting connect from 192.168.0.106)

部署数据服务器

在所有数据服务器上执行以下操作

更新系统

apt update -y && apt upgrade -y && apt install libncurses5 -y

开启防火墙

ufw allow 22ufw allow 2202ufw enable

复制 ndbd 和 ndbmtd 到数据服务器

#复制到192.168.0.106scp ./mysql-cluster-gpl-7.6.17-linux-glibc2.12-x86_64/bin/ndbd ndb1@192.168.0.106:/home/ndb1scp ./mysql-cluster-gpl-7.6.17-linux-glibc2.12-x86_64/bin/ndbmtd ndb1@192.168.0.106:/home/ndb1#复制到192.168.0.104scp ./mysql-cluster-gpl-7.6.17-linux-glibc2.12-x86_64/bin/ndbd ndb2@192.168.0.104:/home/ndb2scp ./mysql-cluster-gpl-7.6.17-linux-glibc2.12-x86_64/bin/ndbmtd ndb2@192.168.0.104:/home/ndb2

在管理服务器复制 ndbd 和 ndbmtd 到/usr/local/bin 文件夹

#192.168.0.106cp -rfv /home/ndb1/ndbd /usr/local/bincp -rfv /home/ndb1/ndbmtd /usr/local/bin#192.168.0.104cp -rfv /home/ndb2/ndbd /usr/local/bincp -rfv /home/ndb2/ndbmtd /usr/local/bin

赋予 ndbd 可执行权限

chmod +x /usr/local/bin/ndbdchmod +x /usr/local/bin/ndbmtd

在/etc下加入my.cnf文件

vi /etc/my.cnf

my.cnf文件

[mysqld]# Options for mysqld process:ndbcluster      # run NDB storage engine[mysql_cluster]# Options for NDB Cluster processes:ndb-connectstring=192.168.0.105 # location of management server

创建数据保存的目录，必须与管理服务配置的路径一致

mkdir -p /data/mysql-cluster/data

启动数据服务

root@ndb1:/usr/local/bin# ndbd2021-06-20 08:10:23 [ndbd] INFO  -- Angel connected to '192.168.0.105:1186'2021-06-20 08:10:23 [ndbd] INFO  -- Angel allocated nodeid: 3

回到集群管理服务器查看集群状态，此时可以看到数据服务已经连接成功

root@mgm:/usr/local/bin# ndb_mgm-- NDB Cluster -- Management Client --ndb_mgm> showConnected to Management Server at: localhost:1186Cluster Configuration---------------------[ndbd(NDB)]	2 node(s)id=2 (not connected, accepting connect from 192.168.0.104)id=3	@192.168.0.106 (mysql-5.7.33 ndb-7.6.17, starting, Nodegroup: 0)[ndb_mgmd(MGM)]	1 node(s)id=1	@192.168.0.105 (mysql-5.7.33 ndb-7.6.17)[mysqld(API)]	2 node(s)id=4 (not connected, accepting connect from 192.168.0.104)id=5 (not connected, accepting connect from 192.168.0.106)

在另一台服务器（192.168.0.104）重复 4、5、6、7 步骤的操作，结果可看到

root@ndb2:/usr/local/bin# ndbd2021-06-20 08:20:10 [ndbd] INFO  -- Angel connected to '192.168.0.105:1186'2021-06-20 08:20:10 [ndbd] INFO  -- Angel allocated nodeid: 2

回到集群管理服务器查看集群状态，此时可以看到所有数据服务已经连接成功

root@mgm:/usr/local/bin# ndb_mgm-- NDB Cluster -- Management Client --ndb_mgm> showConnected to Management Server at: localhost:1186Cluster Configuration---------------------[ndbd(NDB)]	2 node(s)id=2	@192.168.0.104 (mysql-5.7.33 ndb-7.6.17, Nodegroup: 0, *)id=3	@192.168.0.106 (mysql-5.7.33 ndb-7.6.17, Nodegroup: 0)[ndb_mgmd(MGM)]	1 node(s)id=1	@192.168.0.105 (mysql-5.7.33 ndb-7.6.17)[mysqld(API)]	2 node(s)id=4 (not connected, accepting connect from 192.168.0.104)id=5 (not connected, accepting connect from 192.168.0.106)

在目录/data/mysql/data下面可以看到数据服务已经产生了数据

root@ndb1:~# ls /data/mysql/data/ndb_3_fs ndb_3_out.log ndb_3.pid

部署 SQL 服务

复制 MySQL 到SQL服务器

scp ./mysql-cluster-gpl-7.6.17-linux-glibc2.12-x86_64.tar.gz ndb2@192.168.0.104:/home/ndb2scp ./mysql-cluster-gpl-7.6.17-linux-glibc2.12-x86_64.tar.gz ndb1@192.168.0.106:/home/ndb1

解压缩 MySQL, 然后复制到/usr/local目录

tar -zxvf mysql-cluster-gpl-7.6.17-linux-glibc2.12-x86_64.tar.gzcp -rfv mysql-cluster-gpl-7.6.17-linux-glibc2.12-x86_64 /usr/local/ln -snf /usr/local/mysql-cluster-gpl-7.6.17-linux-glibc2.12-x86_64 /usr/local/mysqlcp /usr/local/mysql/support-files/mysql.server /etc/init.d/mysql.serverexport PATH=$PATH:/usr/local/mysql/binsource /etc/profile

开启防火墙

ufw allow 22ufw allow 3306ufw enable

mkdir -p /data/mysql/datamkdir -p /data/mysql/runmkdir -p /var/log/mysql

groupadd mysqluseradd -r -g mysql -s /bin/false mysqlchown mysql:mysql /data/mysql/datachmod 750 /data/mysql/datachown mysql:mysql /data/mysql/runchmod 750 /data/mysql/runchown mysql:mysql /var/log/mysqlchmod 750 /var/log/mysql

创建 MySQL 配置文件

mkdir -p /etc/mysqlvi /etc/mysql/my.cnf

my.cnf

[mysqld]# Options for mysqld process:ndbcluster     # run NDB storage enginepid-file = /data/mysql/run/mysqld.pidsocket = /data/mysql/run/mysqld.sockdatadir = /data/mysql/data# log-error = /var/log/mysql/error.log# By default we only accept connections from localhostbind-address = 192.168.0.106# Disabling symbolic-links is recommended to prevent assorted security riskssymbolic-links = 0[mysql_cluster]# Options for NDB Cluster processes:ndb-connectstring = 192.168.0.105 # location of management server[client]socket = /data/mysql/run/mysqld.sock

初始化MySQL

/usr/local/mysql/bin/mysqld --defaults-file=/etc/mysql/my.cnf --initialize --user=mysql --basedir=/usr/local/mysql

记录下 MySQL 初始化生成的 root 用户密码 sF#Hy,IuT6d#

root@ndb1:~# /usr/local/mysql/bin/mysqld --defaults-file=/etc/mysql/my.cnf --initialize --user=mysql --basedir=/usr/local/mysql2021-06-20T12:23:26.874302Z 0 [Warning] TIMESTAMP with implicit DEFAULT value is deprecated. Please use --explicit_defaults_for_timestamp server option (see documentation for more details).2021-06-20T12:23:27.102146Z 0 [Warning] InnoDB: New log files created, LSN=457902021-06-20T12:23:27.145317Z 0 [Warning] InnoDB: Creating foreign key constraint system tables.2021-06-20T12:23:27.154405Z 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: 50a15854-d1c2-11eb-9792-000c29681e23.2021-06-20T12:23:27.155927Z 0 [Warning] Gtid table is not ready to be used. Table 'mysql.gtid_executed' cannot be opened.2021-06-20T12:23:28.339372Z 0 [Warning] CA certificate ca.pem is self signed.2021-06-20T12:23:28.624534Z 1 [Note] A temporary password is generated for root@localhost: sF#Hy,IuT6d#

启动MySQL

/usr/local/mysql/bin/mysqld_safe --user=mysql &

修改 root 用户密码

mysqladmin -uroot -p'sF#Hy,IuT6d#' password '123456'

回到集群管理服务器查看集群状态，此时可以看到有一个 SQL 服务已经连接上了

root@mgm:/usr/local/bin# ndb_mgm-- NDB Cluster -- Management Client --ndb_mgm> showConnected to Management Server at: localhost:1186Cluster Configuration---------------------[ndbd(NDB)]	2 node(s)id=2	@192.168.0.104 (mysql-5.7.33 ndb-7.6.17, Nodegroup: 0, *)id=3	@192.168.0.106 (mysql-5.7.33 ndb-7.6.17, Nodegroup: 0)[ndb_mgmd(MGM)]	1 node(s)id=1	@192.168.0.105 (mysql-5.7.33 ndb-7.6.17)[mysqld(API)]	2 node(s)id=4 (not connected, accepting connect from 192.168.0.104)id=5	@192.168.0.106 (mysql-5.7.33 ndb-7.6.17)

在另一台服务器（192.168.0.104）部署 SQL 服务，回到集群管理服务器查看集群状态，此时可以看到所有 SQL 服务已经连接成功

root@mgm:/usr/local/bin# ndb_mgm-- NDB Cluster -- Management Client --ndb_mgm> showCluster Configuration---------------------[ndbd(NDB)]	2 node(s)id=2	@192.168.0.104 (mysql-5.7.33 ndb-7.6.17, Nodegroup: 0, *)id=3	@192.168.0.106 (mysql-5.7.33 ndb-7.6.17, Nodegroup: 0)[ndb_mgmd(MGM)]	1 node(s)id=1	@192.168.0.105 (mysql-5.7.33 ndb-7.6.17)[mysqld(API)]	2 node(s)id=4	@192.168.0.104 (mysql-5.7.33 ndb-7.6.17)id=5	@192.168.0.106 (mysql-5.7.33 ndb-7.6.17)

所有集群服务部署完毕，我们来测试一下集群是否真的部署成功

在 192.168.0.106 的 MySQL 上创建数据库和表

CREATE DATABASE `wechat`;CREATE TABLE wechat.user (	Column1 varchar(100) NULL,	Column2 varchar(100) NULL)ENGINE=ndbclusterDEFAULT CHARSET=utf8mb4COLLATE=utf8mb4_general_ci;

插入数据并查看

mysql> show databases;+--------------------+| Database   |+--------------------+| information_schema || mysql    || ndbinfo   || performance_schema || sys    || wechat    |+--------------------+6 rows in set (0.00 sec)mysql> select * from wechat.user;Empty set (0.02 sec)mysql> insert wechat.user (Column1, column2) value ('1', '2');Query OK, 1 row affected (0.01 sec)mysql> select * from wechat.user;+---------+---------+| Column1 | Column2 |+---------+---------+| 1  | 2  |+---------+---------+1 row in set (0.00 sec)

在另一个 SQL 服务器查询，结果是成功的

mysql> show databases;+--------------------+| Database   |+--------------------+| information_schema || mysql    || ndbinfo   || performance_schema || sys    || wechat    |+--------------------+6 rows in set (0.00 sec)mysql> select * from wechat.user;Empty set (0.07 sec)mysql> select * from wechat.user;+---------+---------+| Column1 | Column2 |+---------+---------+| 1  | 2  |+---------+---------+1 row in set (0.00 sec)

现在我们把其中一个数据节点关掉，在管理服务器我们看到 ndbd已经关闭一个了

root@mgm:/usr/local/bin# ndb_mgm-- NDB Cluster -- Management Client --ndb_mgm> showConnected to Management Server at: localhost:1186Cluster Configuration---------------------[ndbd(NDB)]	2 node(s)id=2	@192.168.0.104 (mysql-5.7.33 ndb-7.6.17, Nodegroup: 0, *)id=3 (not connected, accepting connect from 192.168.0.106)[ndb_mgmd(MGM)]	1 node(s)id=1	@192.168.0.105 (mysql-5.7.33 ndb-7.6.17)[mysqld(API)]	2 node(s)id=4	@192.168.0.104 (mysql-5.7.33 ndb-7.6.17)id=5	@192.168.0.106 (mysql-5.7.33 ndb-7.6.17)

写入一笔数据

mysql> select * from wechat.user;+---------+---------+| Column1 | Column2 |+---------+---------+| 1  | 2  |+---------+---------+1 row in set (0.01 sec)mysql> insert into wechat.user (Column1, column2) value ('3', '4');Query OK, 1 row affected (0.00 sec)mysql> select * from wechat.user;+---------+---------+| Column1 | Column2 |+---------+---------+| 3  | 4  || 1  | 2  |+---------+---------+2 rows in set (0.00 sec)

在另一台 SQL 服务器查询，结果还是一致的

mysql> select * from wechat.user;+---------+---------+| Column1 | Column2 |+---------+---------+| 3  | 4  || 1  | 2  |+---------+---------+2 rows in set (0.00 sec)

我们再关闭 192.168.0.106 SQL服务

root@mgm:/usr/local/bin# ndb_mgm-- NDB Cluster -- Management Client --ndb_mgm> showConnected to Management Server at: localhost:1186Cluster Configuration---------------------[ndbd(NDB)]	2 node(s)id=2	@192.168.0.104 (mysql-5.7.33 ndb-7.6.17, Nodegroup: 0, *)id=3 (not connected, accepting connect from 192.168.0.106)[ndb_mgmd(MGM)]	1 node(s)id=1	@192.168.0.105 (mysql-5.7.33 ndb-7.6.17)[mysqld(API)]	2 node(s)id=4	@192.168.0.104 (mysql-5.7.33 ndb-7.6.17)id=5 (not connected, accepting connect from 192.168.0.106)

在 192.168.0.104 的 SQL 服务写入一笔数据

mysql> insert into wechat.user (Column1, column2) value ('5', '6');Query OK, 1 row affected (0.00 sec)mysql> select * from wechat.user;+---------+---------+| Column1 | Column2 |+---------+---------+| 5  | 6  || 3  | 4  || 1  | 2  |+---------+---------+3 rows in set (0.00 sec)

启动 192.168.0.106 的数据服务和SQL服务

root@mgm:/usr/local/bin# ndb_mgm-- NDB Cluster -- Management Client --ndb_mgm> showConnected to Management Server at: localhost:1186Cluster Configuration---------------------[ndbd(NDB)]	2 node(s)id=2	@192.168.0.104 (mysql-5.7.33 ndb-7.6.17, Nodegroup: 0, *)id=3	@192.168.0.106 (mysql-5.7.33 ndb-7.6.17, Nodegroup: 0)[ndb_mgmd(MGM)]	1 node(s)id=1	@192.168.0.105 (mysql-5.7.33 ndb-7.6.17)[mysqld(API)]	2 node(s)id=4	@192.168.0.104 (mysql-5.7.33 ndb-7.6.17)id=5	@192.168.0.106 (mysql-5.7.33 ndb-7.6.17)

在 192.168.0.106 查询数据库发现，发生故障期间产生的数据已经同步了过来

root@ndb1:~# mysql -uroot -pEnter password:Welcome to the MySQL monitor. Commands end with ; or \g.Your MySQL connection id is 4Server version: 5.7.33-ndb-7.6.17-cluster-gpl MySQL Cluster Community Server (GPL)Copyright (c) 2000, 2021, Oracle and/or its affiliates.Oracle is a registered trademark of Oracle Corporation and/or itsaffiliates. Other names may be trademarks of their respectiveowners.Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.mysql> select * from wechat.user;+---------+---------+| Column1 | Column2 |+---------+---------+| 1  | 2  || 5  | 6  || 3  | 4  |+---------+---------+3 rows in set (0.08 sec)

数据库集群部署成功了，总结一下集群的注意事项

创建表的时候，需要设置ENGINE=ndbcluster，具体请看上面的建表脚本。
每个 SQL 服务需要创建一样的用户密码
管理服务器不能全部发生故障，否则集群数据库操作失败。
数据服务器不能全部发生故障，否则集群数据库操作失败。
SQL 服务器发生故障期间建立的数据库，在恢复后不会自动同步新建数据库过来，需要手动在故障恢复后的服务器上创建同名数据库，之后数据才会自动同步过来。
只要管理服务器和数据服务器越多，故障发生时，才能保证数据安全的写入，才不会导致数据库系统不可用。
SQL 服务器越多，把数据库访问的请求通过负载均衡服务分摊到各个 SQL 服务器，才能承受更多的并发量。
集群启动必须按照以下顺序依次启动，管理服务->数据服务->SQL服务。

Electronic News