MySQL高可用工具Orchestrator raft模式部署

Posted 2022-05-17 _雪辉_

文章目录

• • 一、orchestrator介绍
  • 二、orchestrator分布式部署
  • • 2.1 orchestrator元数据库
    • 2.2 安装orchestrator
    • 2.3 修改orchestrator配置文件
    • 2.4 启动orchestrator
    • 2.5 web 管理

一、orchestrator介绍

代码库https://github.com/openark/orchestrator

  Orchestrator是一款开源的mysql复制拓扑管理工具，采用go语言编写，支持MySQL主从复制拓扑关系的调整、支持MySQL主库故障自动切换、手动主从切换等功能。Orchestrator依赖MySQL或者SQLite存储元数据，能够提供Web界面展示MySQL集群的拓扑关系及实例状态，通过Web界面可更改MySQL实例的部分配置信息，同时也提供命令行和api接口，以便更加灵活的自动化运维管理。相比于MHA，Orchestrator更加偏重于复制拓扑关系的管理，能够实现MySQL任一复制拓扑关系的调整，并在此基础上，实现MySQL高可用，另外Orchestrator自身可以部署多个节点，通过raft分布式一致性协议，保证自身的高可用。

二、orchestrator分布式部署

2.1 orchestrator元数据库

数据库部署步骤略

#创建orchestrator用户和库，用户开放主机为orchestrator主机ip
CREATE DATABASE IF NOT EXISTS orchestrator;
CREATE USER 'orchestrator'@'%' IDENTIFIED BY 'orchestrator';
GRANT ALL PRIVILEGES ON `orchestrator`.* TO 'orchestrator'@'%'; 
GRANT SUPER, REPLICATION SLAVE, REPLICATION CLIENT ON *.* TO 'orchestrator'@'%'

目标监控数据库

CREATE USER 'orchestrator'@'%' IDENTIFIED BY 'orchestrator';
GRANT SUPER, PROCESS, REPLICATION SLAVE, RELOAD ON *.* TO 'orchestrator'@'%';
GRANT SELECT ON mysql.slave_master_info TO 'orchestrator'@'%';

2.2 安装orchestrator

yum localinstall orchestrator-3.2.6-1.x86_64.rpm
软件安装在/usr/local/orchestrator/
[root@myrocks-mydb01 ~]# ll /usr/local/orchestrator/
total 19384
-rwxr-xr-x 1 root root 19794824 Jul 27 16:24 orchestrator
-rw-r--r-- 1 root root     7915 Dec 12 19:58 orchestrator.conf.json
-rw-rw-r-- 1 root root     5513 May 24  2021 orchestrator-sample.conf.json
-rw-rw-r-- 1 root root     5100 Jun  4  2020 orchestrator-sample-sqlite.conf.json
drwxr-xr-x 7 root root     4096 Dec 11 18:56 resources

2.3 修改orchestrator配置文件

#仅需更改以下配置，不同节点更改raft bind和raft advertise即可
#配置orchestrator监控的目标数据库信息
  "MySQLTopologyUser": "orchestrator",
  "MySQLTopologyPassword": "orchestrator",
#配置orchestrator后端元数据库信息  
  "MySQLOrchestratorHost": "10.0.213.157",
  "MySQLOrchestratorPort": 33071,
  "MySQLOrchestratorDatabase": "orchestrator",
  "MySQLOrchestratorUser": "orchestrator",
  "MySQLOrchestratorPassword": "orchestrator",
  "RaftEnabled": true,
  "RaftDataDir": "/data/orchestrator",
  "RaftBind": "xxx.xxx.xxx.xxx",
  "RaftAdvertise": "10.0.250.252",
  "DefaultRaftPort": 10008,
  "RaftNodes": [
    "xxx.xxx.xxx.xxx",
    "xxx.xxx.xxx.xxx",
    "xxx.xxx.xxx.xxx"
  ]

2.4 启动orchestrator

启动命令，三个节点都需要启动

cd /usr/local/orchestrator && ./orchestrator --config=./orchestrator.conf.json http 2>&1 &
#三个节点都启动后观察日志可以看到Orchestrator做的事情
2021-12-19 23:29:07 DEBUG Connected to orchestrator backend: orchestrator:?@tcp(10.0.213.157:33071)/orchestrator?timeout=1s&readTimeout=30s&rejectReadOnly=false&interpolateParams=true
2021-12-19 23:29:07 DEBUG Orchestrator pool SetMaxOpenConns: 128
2021-12-19 23:29:07 DEBUG Initializing orchestrator
2021-12-19 23:29:07 INFO Connecting to backend 10.0.213.157:33071: maxConnections: 128, maxIdleConns: 32
2021-12-19 23:29:07 INFO Starting Discovery
2021-12-19 23:29:07 INFO Registering endpoints
2021-12-19 23:29:07 INFO continuous discovery: setting up
2021-12-19 23:29:07 DEBUG Setting up raft
2021-12-19 23:29:07 DEBUG raft: advertise=10.0.250.252:10008
2021-12-19 23:29:07 DEBUG Queue.startMonitoring(DEFAULT)
2021-12-19 23:29:07 DEBUG raft: transport=&connPool:map[] connPoolLock:state:0 sema:0 consumeCh:0xc000224240 heartbeatFn:<nil> heartbeatFnLock:state:0 sema:0 logger:0xc000222dc0 maxPool:3 shutdown:false shutdownCh:0xc0002242a0 shutdownLock:state:0 sema:0 stream:0xc0002200c0 timeout:10000000000 TimeoutScale:262144
2021-12-19 23:29:07 DEBUG raft: peers=[10.0.213.157:10008 10.0.219.146:10008 10.0.250.252:10008]
2021-12-19 23:29:07 DEBUG raft: logStore=&dataDir:/data/orchestrator backend:<nil>
2021-12-19 23:29:07 INFO Starting HTTP listener on :3000
2021-12-19 23:29:07 INFO raft: store initialized at /data/orchestrator/raft_store.db
2021-12-19 23:29:07 INFO new raft created
2021-12-19 23:29:07 INFO continuous discovery: starting
2021/12/19 23:29:07 [INFO] raft: Node at 10.0.250.252:10008 [Follower] entering Follower state (Leader: "")
2021/12/19 23:29:08 [DEBUG] raft-net: 10.0.250.252:10008 accepted connection from: 10.0.213.157:22129
2021/12/19 23:29:08 [DEBUG] raft: Node 10.0.250.252:10008 updated peer set (2): [10.0.213.157:10008 10.0.219.146:10008 10.0.250.252:10008]
2021-12-19 23:29:08 DEBUG orchestrator/raft: applying command 2: leader-uri
2021-12-19 23:29:08 DEBUG orchestrator/raft: applying command 3: request-health-report
2021-12-19 23:29:08 ERROR HttpGetLeader: got 500 status on http://10.0.213.157:3000/api/raft-follower-health-report/1f4a4bb6/10.0.250.252/10.0.250.252
2021-12-19 23:29:08 DEBUG orchestrator/raft: applying command 4: request-health-report
2021-12-19 23:29:08 DEBUG orchestrator/raft: applying command 5: request-health-report
2021-12-19 23:29:08 DEBUG orchestrator/raft: applying command 6: request-health-report
2021-12-19 23:29:08 DEBUG orchestrator/raft: applying command 7: request-health-report
2021-12-19 23:29:08 DEBUG orchestrator/raft: applying command 8: heartbeat
2021-12-19 23:29:08 DEBUG orchestrator/raft: applying command 9: request-health-report
2021-12-19 23:29:08 DEBUG orchestrator/raft: applying command 10: request-health-report

2.5 web 管理

发现节点，主会自动发现从节点，从节点会自动发现主节点

集群拓扑，红色为异常节点