Consul 简介及集群安装

简介

• Consul是基于GO语言开发的开源工具，主要面向分布式，服务化的系统提供服务注册、服务发现和配置管理的功能。
• Consul的功能都很实用，其中包括：服务注册/发现、健康检查、Key/Value存储、多数据中心和分布式一致性保证等特性。
• Consul本身只是一个二进制的可执行文件，所以安装和部署都非常简单，只需要从官网下载后，在执行对应的启动脚本即可。
• 官网下载地址: https://www.consul.io/downloads.html

特性

• 基础特性
  1. 服务注册/发现
  2. 数据强一致型保证
  3. 多数据中心
  4. 健康检查
  5. Key/Value存储
• 高级特性
  • HTTP API
  • ACL

工作模式

• Consul中包括的3种不同的角色：Client、Server、Server-Leader。

• 还有一个在图上没有标出来的角色Agent，一共4个角色。

• Agent

  • 是一个守护线程

  • 跟随Consul应用启动而启动

  • 负责检查、维护节点同步

• Client

  • 转发所有请求给Server
  • 无状态, 不持久化数据
  • 参与LAN Gossip的健康检查

• Server

  • 持久化数据
  • 转发请求给Server-Leader
  • 参与Server-Leader选举
  • 通过WAN Gossip(流言算法协议, Cassandra和比特币的底层算法, 符合弱一致性), 与其他数据中心交换数据

• Server-Leader

  • 响应RPC请求
  • 将服务列表数据同步给Server

工作原理

• 服务注册的方式

  1. HTTP API

     http://{ip}:8500/v1/agent/service/register/:service
     

  2. JSON 配置文件

     {
         "services": [
                 {
                         "id": "serverId",
                         "name": "serverName",
                         "tags": [
                                 "primary"
                         ],
                         "address": "127.0.0.1",
                         "port": 9003,
                         "checks": [
                                 {
                                         "id": "api-servie",
                                         "name": "Service 'xx' check",
                                         "http": "http://127.0.0.1:9003/public/health",
                                         "method": "GET",
                                         "interval": "10s",
                                         "timeout": "1s"
                                 }
                         ]
                 }
         ]
     

     • 启动增加参数: -config-dir

       nohup ./consul agent -dev -config-dir=/consul-conf/service.json &
       

• 服务发现的方式

  1. HTTP API

     • 获取某个service下健康的服务列表信息

       http://{ip}:8500/v1/health/service/:service
       

  2. DNS Agent

• 工作流程

  

集群搭建

• 这次我们以二进制包方式搭建, 其实还有docker的方式。

• 我使用的系统为ubuntu18.04LTS

• 集群规划:node01, node02, node03, node04, node05上都安装

• 集群前置准备请参考: https://www.cnblogs.com/ronnieyuan/p/11461377.html

• 上传二进制包并解压到指定目录

  unzip consul_1.6.2_linux_amd64.zip -d /usr/bin
  

• 解压后直接就是一个可执行文件了

• vim ~/.bashrc 修改环境变量

  # Consul
  export CONSUL_HOME=/usr/bin/consul
  export PATH=$PATH:CONSUL_HOME
  

• 验证Consul安装是否成功

  root@node01:/home/ronnie/soft# consul version
  Consul v1.6.2
  Protocol 2 spoken by default, understands 2 to 3 (agent will automatically use protocol >2 when speaking to compatible agents)
  root@node01:/home/ronnie/soft# vim ~/.bashrc
  

• 将二进制文件拷贝至其他几台虚拟机上, 记得修改配置文件

  cd /usr/bin/
  scp consul root@node02:`pwd`
  scp consul root@node03:`pwd`
  scp consul root@node04:`pwd`
  scp consul root@node05:`pwd`
  

• 分别在五台虚拟机上输入以下对应的命令

  # node01
  consul agent -server -ui -bootstrap-expect=5 -data-dir=/data/consul -node=agent-1 -client=0.0.0.0 -bind=192.168.180.130 -datacenter=dc1
  
  # node02
  consul agent -server -ui -bootstrap-expect=5 -data-dir=/data/consul -node=agent-2 -client=0.0.0.0 -bind=192.168.180.131 -datacenter=dc1 -join 192.168.180.130
  
  # node03
  consul agent -server -ui -bootstrap-expect=5 -data-dir=/data/consul -node=agent-3 -client=0.0.0.0 -bind=192.168.180.132 -datacenter=dc1 -join 192.168.180.130
  
  # node04
  consul agent -server -ui -bootstrap-expect=5 -data-dir=/data/consul -node=agent-4 -client=0.0.0.0 -bind=192.168.180.133 -datacenter=dc1 -join 192.168.180.130
  
  # node05
  consul agent -server -ui -bootstrap-expect=5 -data-dir=/data/consul -node=agent-5 -client=0.0.0.0 -bind=192.168.180.134 -datacenter=dc1 -join 192.168.180.130
  

  • 命令解读:

    • -bootstrap-expect=5: 指定了集群有5台
    • -node=agent-x: 指定服务器
    • -client=0.0.0.0: 指定当前主机客户端侦听地址为0.0.0.0
    • -bind: 绑定当前主机的ip地址
    • -datacenter=dc: 指定了一个数据中心的名称
    • -join 192.168.180.130: 后几台服务器在启动的时候加入第一台代理服务器node01(PS， 比较坑的是即使已经改过hostname文件, 还是得敲ip)
    • -ui指定了启用每台服务器的内置 WebUI 服务器组件
    • 当五台服务器都正确运行起来以后，Consul 集群将自动选举 leader，自动进行集群事务，无需干预。

  • 启动成功案例

    ==> bind_addr cannot be empty
    root@node02:~# consul agent -server -ui -bootstrap-expect=5 -data-dir=/data/consul -node=agent-2 -client=0.0.0.0 -bind=192.168.180.131 -datacenter=dc1 -join 192.168.180.130
    bootstrap_expect > 0: expecting 5 servers
    ==> Starting Consul agent...
               Version: 'v1.6.2'
               Node ID: '304a1a5a-2987-e7ad-738d-7d21f9dbd887'
             Node name: 'agent-2'
            Datacenter: 'dc1' (Segment: '<all>')
                Server: true (Bootstrap: false)
           Client Addr: [0.0.0.0] (HTTP: 8500, HTTPS: -1, gRPC: -1, DNS: 8600)
          Cluster Addr: 192.168.180.131 (LAN: 8301, WAN: 8302)
               Encrypt: Gossip: false, TLS-Outgoing: false, TLS-Incoming: false, Auto-Encrypt-TLS: false
    
    ==> Log data will now stream in as it occurs:
    
        2019/11/29 22:19:15 [INFO]  raft: Initial configuration (index=0): []
        2019/11/29 22:19:15 [INFO]  raft: Node at 192.168.180.131:8300 [Follower] entering Follower state (Leader: "")
        2019/11/29 22:19:15 [INFO] serf: EventMemberJoin: agent-2.dc1 192.168.180.131
        2019/11/29 22:19:15 [INFO] serf: EventMemberJoin: agent-2 192.168.180.131
        2019/11/29 22:19:15 [INFO] consul: Adding LAN server agent-2 (Addr: tcp/192.168.180.131:8300) (DC: dc1)
        2019/11/29 22:19:15 [INFO] agent: Started DNS server 0.0.0.0:8600 (udp)
        2019/11/29 22:19:15 [INFO] consul: Handled member-join event for server "agent-2.dc1" in area "wan"
        2019/11/29 22:19:15 [INFO] agent: Started DNS server 0.0.0.0:8600 (tcp)
        2019/11/29 22:19:15 [INFO] agent: Started HTTP server on [::]:8500 (tcp)
    ==> Joining cluster...
        2019/11/29 22:19:15 [INFO] agent: (LAN) joining: [192.168.180.130]
        2019/11/29 22:19:15 [INFO] serf: EventMemberJoin: agent-1 192.168.180.130
        2019/11/29 22:19:15 [INFO] agent: (LAN) joined: 1
        Join completed. Synced with 1 initial agents
        2019/11/29 22:19:15 [INFO] agent: started state syncer
    ==> Consul agent running!
        2019/11/29 22:19:15 [INFO] consul: Adding LAN server agent-1 (Addr: tcp/192.168.180.130:8300) (DC: dc1)
        2019/11/29 22:19:15 [INFO] serf: EventMemberJoin: agent-1.dc1 192.168.180.130
        2019/11/29 22:19:15 [INFO] consul: Handled member-join event for server "agent-1.dc1" in area "wan"
        2019/11/29 22:19:17 [INFO] serf: EventMemberJoin: agent-3.dc1 192.168.180.132
        2019/11/29 22:19:17 [INFO] consul: Handled member-join event for server "agent-3.dc1" in area "wan"
        2019/11/29 22:19:17 [INFO] serf: EventMemberJoin: agent-3 192.168.180.132
        2019/11/29 22:19:17 [INFO] consul: Adding LAN server agent-3 (Addr: tcp/192.168.180.132:8300) (DC: dc1)
        2019/11/29 22:19:17 [INFO] serf: EventMemberJoin: agent-4.dc1 192.168.180.133
        2019/11/29 22:19:17 [INFO] consul: Handled member-join event for server "agent-4.dc1" in area "wan"
        2019/11/29 22:19:18 [INFO] serf: EventMemberJoin: agent-4 192.168.180.133
        2019/11/29 22:19:18 [INFO] consul: Adding LAN server agent-4 (Addr: tcp/192.168.180.133:8300) (DC: dc1)
        2019/11/29 22:19:18 [INFO] serf: EventMemberJoin: agent-5 192.168.180.134
        2019/11/29 22:19:18 [INFO] consul: Adding LAN server agent-5 (Addr: tcp/192.168.180.134:8300) (DC: dc1)
        2019/11/29 22:19:18 [INFO] consul: Existing Raft peers reported by agent-5, disabling bootstrap mode
        2019/11/29 22:19:18 [INFO] serf: EventMemberJoin: agent-5.dc1 192.168.180.134
        2019/11/29 22:19:18 [INFO] consul: Handled member-join event for server "agent-5.dc1" in area "wan"
        2019/11/29 22:19:19 [WARN]  raft: Failed to get previous log: 1 log not found (last: 0)
        2019/11/29 22:19:19 [INFO] agent: Synced node info
        2019/11/29 22:19:19 [INFO] consul: New leader elected: agent-1
    ==> Failed to check for updates: Get https://checkpoint-api.hashicorp.com/v1/check/consul?arch=amd64&os=linux&signature=6977bf57-e750-2607-d65d-fa3ab1444055&version=1.6.2: net/http: request canceled (Client.Timeout exceeded while awaiting headers)
    

• 在浏览器上访问node01节点的8500端口

  

  

  

• 那么本次集群的安装就到此为止了