环境
OS:Centos 7
mysql(5.7)主从:
192.168.1.100 hmaster 主
192.168.1.102 hslavea 从
192.168.1.103 hslaveb 从
192.168.1.104 orchestrator元数据使用的数据库
前提条件:部署好被管控的mysql主从
1.下载安装介质
下载地址:https://github.com/openark/orchestrator/releases,选择相应的版本
orchestrator-client-3.1.4-1.x86_64.rpm
orchestrator-3.1.4-1.x86_64.rpm
下面两个安装包需要自行下载
oniguruma-6.8.2-2.el7.x86_64.rpm
jq-1.6-2.el7.x86_64.rpm
2.安装在每个mysql服务器上都进行安装
rpm -ivh orchestrator-client-3.1.4-1.x86_64.rpm
rpm -ivh orchestrator-3.1.4-1.x86_64.rpm
rpm -ivh oniguruma-6.8.2-2.el7.x86_64.rpm
rpm -ivh jq-1.6-2.el7.x86_64.rpm
3.设置hostname
每台机器设置相应的hostname
hostnamectl set-hostname hmaster ##192.168.1.100
hostnamectl set-hostname hslavea ##192.168.1.102
hostnamectl set-hostname hslaveb ##192.168.1.103
hostnamectl set-hostname orcmgr ##192.168.1.104
4.设置hostname
每台机器设置hosts文件
vi /etc/hosts文件添加如下项目
192.168.1.100 hmaster
192.168.1.102 hslavea
192.168.1.103 hslaveb
192.168.1.104 orcmgr
5.orchestrator元数据使用的数据库安装
192.168.1.104上操作
安装mysql数据库步骤省略,然后需要创建元数据库和用户
[mysql@localhost bin]$ /opt/mysql57/bin/mysql -h localhost -uroot -p -S /opt/mysql57/mysql.sock
mysql>create database orchestrator;
mysql>create user 'orchestrator'@'%' identified by 'mysql';
mysql>grant all on orchestrator.* to 'orchestrator'@'%';
6.在主库上创建用户
192.168.1.100上操作
GRANT SELECT, RELOAD, PROCESS, SUPER, REPLICATION SLAVE, REPLICATION CLIENT ON *.* TO 'orchestrator'@'%' identified by 'mysql';
创建后会自动同步到另外2个从库
7.编辑配置文件
在192.168.1.100上操作,然后拷贝到另外2个节点
拷贝一个模板的过来进行修改
[root@localhost orchestrator]# cd /usr/local/orchestrator/
[root@localhost orchestrator]# cp orchestrator-sample.conf.json orchestrator.conf.json
{
"Debug": true,
"EnableSyslog": false,
"ListenAddress": ":3000",
"MySQLTopologyUser": "orchestrator",
"MySQLTopologyPassword": "mysql",
"MySQLTopologyCredentialsConfigFile": "",
"MySQLTopologySSLPrivateKeyFile": "",
"MySQLTopologySSLCertFile": "",
"MySQLTopologySSLCAFile": "",
"MySQLTopologySSLSkipVerify": true,
"MySQLTopologyUseMutualTLS": false,
"MySQLOrchestratorHost": "192.168.1.104",
"MySQLOrchestratorPort": 3306,
"MySQLOrchestratorDatabase": "orchestrator",
"MySQLOrchestratorUser": "orchestrator",
"MySQLOrchestratorPassword": "mysql",
"MySQLOrchestratorCredentialsConfigFile": "",
"MySQLOrchestratorSSLPrivateKeyFile": "",
"MySQLOrchestratorSSLCertFile": "",
"MySQLOrchestratorSSLCAFile": "",
"MySQLOrchestratorSSLSkipVerify": true,
"MySQLOrchestratorUseMutualTLS": false,
"MySQLConnectTimeoutSeconds": 1,
"RaftEnabled": true,
"RaftDataDir": "/data/orchestrator",
"RaftBind": "192.168.1.100",
"DefaultRaftPort": 10008,
"RaftNodes": [
"192.168.1.100",
"192.168.1.102",
"192.168.1.103"
],
模板默认是没有raft配置项的,需要自行加入
"RaftEnabled": true,
"RaftDataDir": "/data/orchestrator",
"RaftBind": "192.168.1.100",
"DefaultRaftPort": 10008,
"RaftNodes": [
"192.168.1.100",
"192.168.1.102",
"192.168.1.103"
],
需要创建目录(每个数据库节点都需要执行)
mkdir -p /data/orchestrator
8.拷贝配置文件到另外2个节点
scp /usr/local/orchestrator/orchestrator.conf.json root@192.168.1.102:/usr/local/orchestrator/
scp /usr/local/orchestrator/orchestrator.conf.json root@192.168.1.103:/usr/local/orchestrator/
拷贝过来修改的地方:
“RaftBind”: “192.168.1.100” 这里分别修改更各自节点的ip
9.启动
每台机器都需要执行,也可以使用nohup后台执行
cd /usr/local/orchestrator
./orchestrator –config=./orchestrator.conf.json http
10.web访问
每个数据库节点都可以访问
http://192.168.1.100:3000
http://192.168.1.102:3000
http://192.168.1.103:3000
1.主从切换
orchestrator-client -c graceful-master-takeover -a hslavea:3306 -d hmaster:3306
新主库为:hmaster,但是hslavea不会自动成为新主库的从,需要手工指定后重启动
mysql> CHANGE MASTER TO MASTER_HOST='hmaster',
MASTER_PORT=3306,
MASTER_USER='repl',
MASTER_PASSWORD='mysql',
MASTER_AUTO_POSITION=1,
MASTER_CONNECT_RETRY=1,
MASTER_RETRY_COUNT=86400,
MASTER_HEARTBEAT_PERIOD=2;
mysql> start slave;
2.查看拓扑情况
[root@hmaster orchestrator]# orchestrator-client -c topology -i hmaster:3306
hmaster:3306 [0s,ok,5.7.29-log,rw,ROW,>>,GTID]
+ hslavea:3306 [0s,ok,5.7.29-log,ro,ROW,>>,GTID]
+ hslaveb:3306 [0s,ok,5.7.29-log,rw,ROW,>>,GTID]
— The End–
原文地址：http://www.cnblogs.com/hxlasky/p/16858590.html
