1、準(zhǔn)備三臺(tái)服務(wù)器
2、為三臺(tái)機(jī)器分別安裝Linux操作系統(tǒng)(Oracle Linux / RHEL 6.5 x86_64bit)
3、分別IP地址
管理節(jié)點(diǎn)      192.168.1.110        (負(fù)責(zé)管理整個(gè)集群)
SQL節(jié)點(diǎn)      192.168.1.111        (負(fù)責(zé)操作數(shù)據(jù)庫)
SQL節(jié)點(diǎn)      192.168.1.112        (負(fù)責(zé)操作數(shù)據(jù)庫)
數(shù)據(jù)節(jié)點(diǎn)      192.168.1.111        (負(fù)責(zé)存儲(chǔ)數(shù)據(jù))
數(shù)據(jù)節(jié)點(diǎn)      192.168.1.112        (負(fù)責(zé)存儲(chǔ)數(shù)據(jù))
 
SQL節(jié)點(diǎn)和數(shù)據(jù)節(jié)點(diǎn)可以同在一臺(tái)機(jī)器上
 
4、修改三臺(tái)服務(wù)器的/etc/hosts文件，修改完后的內(nèi)容如下:
127.0.0.1    localhost    localhost.domain
192.168.1.110    node01
192.168.1.111    node02
192.168.1.112    node03
 
5、修改三臺(tái)服務(wù)器的/etc/sysconfig/nework文件，修改完后的內(nèi)容分別如下：

6、分別重啟三臺(tái)服務(wù)器

7、下載NDB集群軟件mysql-cluster-advanced-7.3.5-linux-glibc2.5-x86_64.tar.gz

集群軟件有三種格式：
a、以tar.gz結(jié)尾的二制進(jìn)格式
b、以rpm結(jié)尾的rpm包形式
c、以源文件編譯安裝方式
 
rpm包安裝方式，由于是自動(dòng)安裝的，所以不方便配置軟件安裝位置和數(shù)據(jù)存儲(chǔ)位置
 
所以通常選用.tar.gz結(jié)尾的二進(jìn)制格式的安裝文件
 
注意：mysql-cluster-advanced-7.3.5-linux-glibc2.5-x86_64.tar.gz包含了mysql數(shù)據(jù)庫軟件和NDB集群軟件
 
8、分別復(fù)制mysql-cluster-advanced-7.3.5-linux-glibc2.5-x86_64.tar.gz文件到三臺(tái)服務(wù)器上的/usr/local/目錄下
 
9、新建用戶mysql
# groupadd mysql
# useradd -r -g mysql mysql      (由于mysql用戶只是用于安裝配置NDB集群或mysql軟件，而不用用于登陸系統(tǒng))
    (所以以-r參數(shù)來創(chuàng)建mysql用戶，并且不用為該用戶設(shè)置密碼)
 
10、分別在三臺(tái)服務(wù)器上解壓mysql-cluster-advanced-7.3.5-linux-glibc2.5-x86_64.tar.gz文件

配置完成以后，結(jié)果如下

11、在192.168.1.111和192.168.1.112兩個(gè)數(shù)據(jù)節(jié)點(diǎn)(兩臺(tái)服務(wù)器)上安裝mysql數(shù)據(jù)庫軟件
由于是二進(jìn)制安裝，所以mysql數(shù)據(jù)庫軟件就安裝在/usr/local/mysql下，
    但數(shù)據(jù)庫的數(shù)據(jù)可以存放到其它地方，如/u01/app/mysql/data/下面
 
注意：以下命令在兩臺(tái)數(shù)據(jù)節(jié)點(diǎn)服務(wù)器上都要執(zhí)行

# mkdir -p /u01/app/mysql/data
# chown -R mysql.mysql /u01
# cd /usr/local/mysql
# scripts/mysql_install_db --basedir=/usr/local/mysql  --datadir=/u01/app/mysql/data --user=mysql
執(zhí)行完這條命令以后，數(shù)據(jù)庫的數(shù)據(jù)文件(包括mysql，test , performance_schema等數(shù)據(jù)庫)就被安裝到了
/u01/app/mysql/data目錄下了
 
# cp bin/ndbd  /usr/local/bin/
# cp bin/ndbmtd /usr/local/bin/

然后將啟動(dòng)數(shù)據(jù)庫服務(wù)的啟動(dòng)文件復(fù)制到/etc/rc.d/init.d/目錄下
 
# cp support-files/mysql.server  /etc/rc.d/init.d/mysqld    (改名)
執(zhí)行完上面這條命令以后，就可以使用service mysqld start | stop | restart命令來管理數(shù)據(jù)庫服務(wù)了
 
# cp my.cnf  /etc/my.cnf
# vi /etc/my.cnf
[mysqld]
ndbcluster
basedir=/usr/local/mysql
datadir=/u01/app/mysql/data
port=3306
sql_mode=NO_ENGINE_SUBSTITUTION,STRICT_TRANS_TABLES
 
[mysql_cluster]
ndb-connectstring=192.168.1.110
 
再次注意：以上命令在兩臺(tái)數(shù)據(jù)節(jié)點(diǎn)服務(wù)器上都要執(zhí)行
 
到此，數(shù)據(jù)節(jié)點(diǎn)和SQL節(jié)點(diǎn)的安裝配置就結(jié)束了，如果數(shù)據(jù)節(jié)點(diǎn)和SQL節(jié)點(diǎn)是分開在不同的機(jī)器上，則每個(gè)數(shù)據(jù)節(jié)點(diǎn)
和每個(gè)SQL節(jié)點(diǎn)都要執(zhí)行以上命令。
 
12、在管理節(jié)點(diǎn)上進(jìn)行安裝配置
 
由于第10步已經(jīng)解壓縮mysql-cluster-advanced-7.3.5-linux-glibc2.5-x86_64.tar.gz 安裝包到了/usr/local
目錄下，并進(jìn)行了軟連接等操作，所以這些工作就不用重復(fù)做了，接著往下配置即可
 
# cd /usr/local/mysql
# mkdir mysql-cluster    (創(chuàng)建mysql-cluster目錄，后面啟動(dòng)集群管理服務(wù)時(shí)要用)
# cp bin/ndb_mgm*    /usr/local/bin/
# cd /var/lib
# mkdir mysql-cluster
# cd mysql-cluster
# vi config.ini  (這是管理節(jié)點(diǎn)上的集群配置文件，很重要)
[ndbd default]
NoOfReplicas=1    # Number of replicas
DataMemory=80M    # How much memory to allocate for data storage
IndexMemory=18M  # How much memory to allocate for index storage
                # For DataMemory and IndexMemory, we have used the
                  # default values. Since the "world" database takes up
                  # only about 500KB, this should be more than enough for
                  # this example Cluster setup.
 
[tcp default]
                  # This the default; however, you can use any
                  # port that is free for all the hosts in the cluster
                  # Note: It is recommended that you do not specify the port
                  # number at all and simply allow the default value to be used
                  # instead
 
[ndb_mgmd]
hostname=192.168.1.110          # Hostname or IP address of MGM node
datadir=/var/lib/mysql-cluster  # Directory for MGM node log files
NodeId=1
 
[ndbd]
hostname=192.168.1.111        # Hostname or IP address
datadir=/u01/app/mysql/data  # Directory for this data node's data files
NodeId=2
 
[ndbd]
hostname=192.168.1.112        # Hostname or IP address
datadir=/u01/app/mysql/data  # Directory for this data node's data files
NodeId=3
 
 
[mysqld]
hostname=192.168.1.111        # Hostname or IP address
NodeId=4                      # (additional mysqld connections can be
                              # specified for this node for various
                              # purposes such as running ndb_restore)
 
[mysqld]
hostname=192.168.1.112        # Hostname or IP address
NodeId=5                      # (additional mysqld connections can be
                              # specified for this node for various
                              # purposes such as running ndb_restore)
 
配置說明：
[ndbd default]                這部分是公共部分，對(duì)于每一個(gè)數(shù)據(jù)節(jié)點(diǎn)都有效，只需要配置一份
NoOfReplicas=1                數(shù)據(jù)鏡像幾份(各數(shù)據(jù)節(jié)點(diǎn)之間相互備份)
 
[tcp default]                針對(duì)每個(gè)數(shù)據(jù)節(jié)點(diǎn)及管理節(jié)點(diǎn)之間使用哪個(gè)端口進(jìn)行通訊，
                              在舊版本的NDB集群軟件配置時(shí)，這個(gè)地方通常配置portnumber=2202
                              但新版的NDB軟件這里不需要配置，并且MySQL官方也強(qiáng)烈建議不要配置
 
[ndb_mgmd]                    管理節(jié)點(diǎn)的配置部分(通常只有一個(gè))
                              注意NodeId=1指明管理節(jié)點(diǎn)的節(jié)點(diǎn)ID為1，如果不指定，在啟動(dòng)集群時(shí)，會(huì)報(bào)錯(cuò)
hostname=192.168.1.110          指明管理節(jié)點(diǎn)的IP地址
datadir=/var/lib/mysql-cluster  指明集群管理日志存放的位置
 
[ndbd]                        數(shù)據(jù)節(jié)點(diǎn)配置部分，有幾個(gè)數(shù)據(jù)節(jié)點(diǎn)就配置幾個(gè)[ndbd]
hostname=192.168.1.111        指明數(shù)據(jù)節(jié)點(diǎn)的IP地址
datadir=/u01/app/mysql/data    指明數(shù)據(jù)節(jié)點(diǎn)上的數(shù)據(jù)庫文件存放的位置
NodeId=2                      指明該數(shù)據(jù)節(jié)點(diǎn)在整個(gè)集群中的nodeid號(hào)(很重要)
 
[mysqld]                      SQL節(jié)點(diǎn)配置部分，有幾個(gè)SQL節(jié)點(diǎn)，就配置幾個(gè)[mysqld]

13、配置好上面的所有配置以后就可以在管理節(jié)點(diǎn)上啟動(dòng)怎個(gè)集群了
 
注意： 在管理節(jié)點(diǎn)上，不用執(zhí)行/usr/local/mysql/scripts/mysql_install_db --basedir= --datadir= --user=命令
來安裝數(shù)據(jù)庫了。
 
# ndb_mgmd -f /var/lib/mysql-cluster/config.ini (第一次啟動(dòng)時(shí)這樣執(zhí)行，如果后面新添加了數(shù)據(jù)節(jié)點(diǎn))
(執(zhí)行此命令時(shí)，需要帶上--initital參數(shù)，否則新添加的節(jié)點(diǎn))
(無法被識(shí)別)
(# ndb_mgmd -f /var/lib/mysql-cluster/config.ini --initial)
 
# ndb_mgm (查看啟動(dòng)后的集群狀態(tài)，看看集群是否成功啟動(dòng)，如果看到以下內(nèi)容，表示集群已經(jīng)成功配置并啟動(dòng))
 
[root@mysql01 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 (not connected, accepting connect from 192.168.1.111)
id=3 (not connected, accepting connect from 192.168.1.112)
 
[ndb_mgmd(MGM)] 1 node(s)
id=1 @192.168.1.110 (mysql-5.6.17 ndb-7.3.5)
 
[mysqld(API)] 2 node(s)
id=4 (not connected, accepting connect from 192.168.1.111)
id=5 (not connected, accepting connect from 192.168.1.112)
 
現(xiàn)在看到集群中包含兩個(gè)數(shù)據(jù)節(jié)點(diǎn)、一個(gè)管理節(jié)點(diǎn)、兩個(gè)SQL節(jié)點(diǎn)，但是數(shù)據(jù)節(jié)點(diǎn)和SQL節(jié)點(diǎn)都還沒有啟動(dòng)，
所以顯示的連接狀態(tài)是not connected
 
14、分別登陸兩臺(tái)數(shù)據(jù)節(jié)點(diǎn)執(zhí)行以下命令啟動(dòng)數(shù)據(jù)節(jié)點(diǎn)
# ndbd --initial (第一次啟動(dòng)時(shí)，需要加--initial來初始化數(shù)據(jù)節(jié)點(diǎn)，第二次啟動(dòng)時(shí)，就不需要這個(gè)參數(shù)了)
 
[root@mysql02 support-files]# ndbd --initial
2014-06-12 05:42:04 [ndbd] INFO -- Angel connected to '192.168.1.110:1186'
2014-06-12 05:42:04 [ndbd] INFO -- Angel allocated nodeid: 2
 
 
[root@mysql03 bin]# ndbd --initial
2014-06-12 05:41:38 [ndbd] INFO -- Angel connected to '192.168.1.110:1186'
2014-06-12 05:41:38 [ndbd] INFO -- Angel allocated nodeid: 3
 
15、再登陸到管理節(jié)點(diǎn)查看現(xiàn)在整個(gè)集群的狀態(tài)
ndb_mgm> show
Cluster Configuration
---------------------
[ndbd(NDB)] 2 node(s)
id=2 @192.168.1.111 (mysql-5.6.17 ndb-7.3.5, Nodegroup: 0, *)
id=3 @192.168.1.112 (mysql-5.6.17 ndb-7.3.5, Nodegroup: 1)
 
[ndb_mgmd(MGM)] 1 node(s)
id=1 @192.168.1.110 (mysql-5.6.17 ndb-7.3.5)
 
[mysqld(API)] 2 node(s)
id=4 (not connected, accepting connect from 192.168.1.111)
id=5 (not connected, accepting connect from 192.168.1.112)
 
現(xiàn)在可以看到，兩個(gè)數(shù)據(jù)節(jié)點(diǎn)已經(jīng)連接上來了，表示兩個(gè)數(shù)據(jù)節(jié)點(diǎn)成功啟動(dòng)了
 
16、分別登陸到兩個(gè)SQL節(jié)點(diǎn)上啟動(dòng)SQL節(jié)點(diǎn)
[root@mysql02 ~]# cd /usr/local/mysql/bin
[root@mysql02 bin]# ./mysqld_safe --user=mysql
140612 05:51:00 mysqld_safe Logging to '/u01/app/mysql/data/mysql02.err'.
140612 05:51:00 mysqld_safe Starting mysqld daemon with databases from /u01/app/mysql/data
 
[root@mysql03 ~]# cd /usr/local/mysql/bin
[root@mysql03 bin]# ./mysqld_safe --user=mysql
140612 05:52:07 mysqld_safe Logging to '/u01/app/mysql/data/mysql03.err'.
140612 05:52:07 mysqld_safe Starting mysqld daemon with databases from /u01/app/mysql/data
 
17、再次回到管理節(jié)點(diǎn)查看整個(gè)集群的狀態(tài)
ndb_mgm> show
Cluster Configuration
---------------------
[ndbd(NDB)] 2 node(s)
id=2 @192.168.1.111 (mysql-5.6.17 ndb-7.3.5, Nodegroup: 0, *)
id=3 @192.168.1.112 (mysql-5.6.17 ndb-7.3.5, Nodegroup: 1)
 
[ndb_mgmd(MGM)] 1 node(s)
id=1 @192.168.1.110 (mysql-5.6.17 ndb-7.3.5)
 
[mysqld(API)] 2 node(s)
id=4 @192.168.1.111 (mysql-5.6.17 ndb-7.3.5)
id=5 @192.168.1.112 (mysql-5.6.17 ndb-7.3.5)
 
到此整個(gè)集群就搭建完成，并成功啟動(dòng)運(yùn)行了
 
18、安全關(guān)閉整個(gè)集群
首先登陸到兩個(gè)SQL節(jié)點(diǎn)，執(zhí)行service mysqld stop命令關(guān)閉SQL節(jié)點(diǎn)
[root@mysql02 ~]# service mysqld stop
Shutting down MySQL...... SUCCESS!
 
[root@mysql03 ~]# service mysqld stop
Shutting down MySQL..... SUCCESS!
 
然后登陸到管理節(jié)點(diǎn)上，執(zhí)行shutdown命令關(guān)閉整個(gè)集群
[root@mysql01 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 (not connected, accepting connect from 192.168.1.111)
id=3 (not connected, accepting connect from 192.168.1.112)
 
[ndb_mgmd(MGM)] 1 node(s)
id=1 @192.168.1.110 (mysql-5.6.17 ndb-7.3.5)
 
[mysqld(API)] 2 node(s)
id=4 (not connected, accepting connect from 192.168.1.111)
id=5 (not connected, accepting connect from 192.168.1.112)
 
ndb_mgm> shutdown
1 NDB Cluster node(s) have shutdown.
Disconnecting to allow management server to shutdown.
ndb_mgm> exit
 
 
7 測試。
 
1、從SQL節(jié)點(diǎn)A登錄，創(chuàng)建數(shù)據(jù)庫和表，進(jìn)行簡單測試。
 
mysql> create database zxztest ;
 
mysql> use zxztest;
Database changed
 
mysql> create table test1(id int,name varchar(10)) engine=ndb ;
 
mysql> insert into test1 values(1,'linuxidc');
 
mysql> select * from test1 ;
 
+------+---------+
| id | name |
+------+---------+
| 1 | linuxidc |
+------+---------+

最新評(píng)論