Mysql8 NDB cluster installation and deployment

1、NDB Cluster Introduce

MySQL NDB Cluster Use MySQL The server and NDB Storage engine ,NDBCLUSTER （ Also known as NDB） It's a kind of Memory storage engine , Provide high availability and data persistence .NDB Cluster take MySQL The server and NDB The memory cluster storage engine is integrated together .NDB The cluster consists of a set of hosts , Each host runs one or more processes . These processes, called nodes, can include MySQL The server （ Used to access the NDB data ）、 Data nodes （ For storing data ）、 One or more management servers .

（1） The management node （Management node）： This type of node is used to manage NDB Cluster Other nodes in , Execute provide configuration data 、 Start and stop the node and run backup . Because this node type manages the configuration of other nodes , So you should start this type of node first , Then start any other nodes . Use command ndb_mgmd Start the management node

（2） Data nodes （Data node）：NDB Cluster Core functions , Used to save data 、 Indexes , Control the transaction . The inserted data is stored in different node groups according to the hash value of the primary key （ Each node group holds some data ）, In addition, within each node group , Data is replicated to different data nodes for redundancy

（3）SQL node （SQL node）： It's equivalent to an increase in NDB Storage engine MySQL The server , Data nodes act as NDB Storage engine use , If you use other MySQL Storage engine , for example InnoDB,MyISAM etc. , The data will be saved in SQL Node ; Application through SQL Nodes access data nodes , The method of use is the same as the usual MySQL equally ,SQL The node will automatically find the correct data node to retrieve the data

2、 Environmental planning

Node type	IP	Host name	Occupied port
The management node	172.16.80.56	mgr0	1186
Data nodes	172.16.80.57 172.16.80.58	mgr1 mgr2	2202
SQL node	172.16.80.59	mgr3	3306,33060

3、 download MySQL NDB Cluster

Community Edition ：MySQL :: Download MySQL Cluster

Business Edition ：https://edelivery.oracle.com/osdc/faces/SoftwareDelivery

choice glibc Binary installation package ：

4、 All nodes are initialized

（1） Uninstall existing mysql client

rpm -qa|grep mysql rpm -qa|grep mariadb rpm -e *** uninstall

（2） establish mysql Catalog

lvcreate -L 300G -n lvmysql vgtest mkfs.xfs /dev/vgtest/lvmysql mount /dev/vgtest/lvmysql /mysql

（3） establish mysql Users and groups

groupadd -g 512 mysql useradd -u 512 -g 512 -d /mysql mysql chown -R mysql.mysql /mysql

（4） decompression glibc Package to /mysql

su - mysql cd /mysql tar -xvf /soft/mysql-cluster-commercial-8.0.29-linux-glibc2.12-x86_64.tar.gz mv mysql-cluster-commercial-8.0.29-linux-glibc2.12-x86_64 mysql-cluster-8.0.29 mkdir /mysql/data

（5） Configure environment variables

vi /etc/profile export MYSQL_CLUSTER_HOME=/mysql/mysql-cluster-8.0.29 export PATH="$PATH:${MYSQL_CLUSTER_HOME}/bin"

5、 Manage node configuration

（1） To configure config.ini

cat /mysql/config.ini [ndbd default] # Options affecting ndbd processes on all data nodes: NoOfReplicas=2 # Number of fragment replicas, Specifies the number of copies of each piece of data that are redundantly stored on different nodes , # This parameter is one General should at least be set to 2, The number of segments is the total number of data nodes divided by NoOfReplicas DataMemory=98M # How much memory to allocate for data storage BackupDataDir=/mysql/backup # Data node backup directory [ndb_mgmd] # The management node : HostName=172.16.80.56 DataDir=/mysql/data [ndbd] # Data nodes "A": HostName=172.16.80.57 NodeId=2 DataDir=/mysql/data [ndbd] # Data nodes "B": HostName=172.16.80.58 NodeId=3 DataDir=/mysql/data [mysqld] # SQL node : HostName=172.16.80.59

（2） To configure my.cnf

cat /mysql/my.cnf [mysqld] # Options for mysqld process: ndbcluster # run NDB storage engine datadir=/mysql/data basedir=/mysql/mysql-cluster-8.0.29 [mysql_cluster] # Options for NDB Cluster processes: ndb-connectstring=172.16.80.56 # location of management server

（3） Initialize and start the management node

ndb_mgmd --defaults-file=/mysql/my.cnf --initial --config-file /mysql/config.ini --configdir=/mysql

（4） Restart the management node and view the cluster status

ndb_mgm show shutdown exit Start the management node normally ： ndb_mgmd --defaults-file=/mysql/my.cnf --config-file /mysql/config.ini --configdir=/mysql

6、 Data node configuration （ Both data nodes execute ）

（1） To configure my.cnf

cat /mysql/my.cnf [mysql_cluster] # Options for NDB Cluster processes: ndb-connectstring=172.16.80.56 # location of management server

（2） Start two ndbd Data nodes

ndbd --defaults-file=/mysql/my.cnf

7、MySQL Node configuration

（1） Create the necessary Directory

su - mysql mkdir -p /mysql/log

（2） establish my.cnf

cat /mysql/my.cnf [mysqld] # Options for mysqld process: ndbcluster # run NDB storage engine datadir=/mysql/data basedir=/mysql/mysql-cluster-8.0.29 log-error=/mysql/log/mysqld.log pid-file=/mysql/mysqld.pid [mysql_cluster] # Options for NDB Cluster processes: ndb-connectstring=172.16.80.56 # location of management server

（3） establish /etc/my.cnf Soft link

ln -s /mysql/my.cnf /etc/my.cnf

（4） establish mysqld Service and start

cp /mysql/mysql-cluster-8.0.29/support-files/mysql.server /etc/init.d/mysqld service mysqld status service mysqld start

（5） Reset root password

Check the temporary password /mysql/log/mysqld.log

mysql -uroot -p alter user [email protected] identified by 'R00t_123'; exit mysql -uroot -pR00t_123 show engines;

8、 Start stop test

Stop sequence ：

SQL node --> Data nodes --> The management node

Startup sequence ：

The management node --> Data nodes -->SQL node

Stop the order ：

service mysqld stop（SQL node ）

ndb_mgm -e shutdown（ The management node ）

Start command ：

ndb_mgmd --defaults-file=/mysql/my.cnf --config-file /mysql/config.ini --configdir=/mysql（ The management node ）

ndbd --defaults-file=/mysql/my.cnf （ All data nodes ）

service mysqld start（SQL node ）

9、 establish NDB Table test

（1） establish InnoDB surface

create database ndb_test; use ndb_test; CREATE TABLE `test1` ( `id` int NOT NULL AUTO_INCREMENT, `name` varchar(10), PRIMARY KEY (`id`) ) ENGINE=InnoDB;

Find out InnoDB Tables are built on SQL nodes

（2） establish ndb surface

CREATE TABLE `test2` ( `id` int NOT NULL AUTO_INCREMENT, `name` varchar(10) , PRIMARY KEY (`id`) ) ENGINE=ndbcluster/ndb;

（3） Test both tables insert/select Speed

DROP PROCEDURE IF EXISTS repeatCourse; DELIMITER $ CREATE PROCEDURE repeatCourse() BEGIN DECLARE i INT DEFAULT 1; WHILE i < 400000 DO insert into ndb_test.test2 values(null,i); SET i = i + 1; END WHILE; END $ DELIMITER ; CALL repeatCourse(); It seems to be ndb Tables are faster