Zookeeper 集群模式

2021 年 11 月 18 日

922 次浏览

4061字数

一个 zooKeeper 集群通常由3台及以上的 zooKeeper 服务组成；组成 zooKeeper 集群的每台机器都会在内存中维护当前的服务器状态，并且每台机器之间都会互相保持通信。只要集群中存在超过一半的机器能够正常工作，那么整个集群就能够正常对外服务。

Zookeeper 集群角色：

Leader：处理所有的事务请求，处理读请求，集群中只能有一个 Leader。
Follower：处理读请求，并转发事务请求给 Leader 服务器；参与事务请求的同步提交投票，与 Leader 进行数据交换；参与 Leader选举投票。
Observer：处理读请求；不能参与选举；不影响集群写性能的情况下，提高集群读性能。

Zookeeper 集群搭建：

使用docker-compose配置zoo1～zoo4，其中zoo4配置为 observer 服务；创建对应的data、datalog文件夹。

version: '3.7'

networks:
  docker_net:
    external: true

services:
  zoo1:
    image: zookeeper:3.6.3
    restart: unless-stopped
    hostname: zoo1
    container_name: zoo1
    ports:
      - 2182:2181
    environment:
      ZOO_MY_ID: 1
      ZOO_SERVERS: server.1=0.0.0.0:2888:3888;2181 server.2=zoo2:2888:3888;2181 server.3=zoo3:2888:3888;2181 server.4=zoo4:2888:3888:observer;2181
    volumes:
      - ./zoo1/data:/data
      - ./zoo1/datalog:/datalog
    networks:
      - docker_net

  zoo2:
    image: zookeeper:3.6.3
    restart: unless-stopped
    hostname: zoo2
    container_name: zoo2
    ports:
      - 2183:2181
    environment:
      ZOO_MY_ID: 2
      ZOO_SERVERS: server.1=zoo1:2888:3888;2181 server.2=0.0.0.0:2888:3888;2181 server.3=zoo3:2888:3888;2181 server.4=zoo4:2888:3888:observer;2181
    volumes:
      - ./zoo2/data:/data
      - ./zoo2/datalog:/datalog
    networks:
      - docker_net

  zoo3:
    image: zookeeper:3.6.3
    restart: unless-stopped
    hostname: zoo3
    container_name: zoo3
    ports:
      - 2184:2181
    environment:
      ZOO_MY_ID: 3
      ZOO_SERVERS: server.1=zoo1:2888:3888;2181 server.2=zoo2:2888:3888;2181 server.3=0.0.0.0:2888:3888;2181 server.4=zoo4:2888:3888:observer;2181
    volumes:
      - ./zoo3/data:/data
      - ./zoo3/datalog:/datalog
    networks:
      - docker_net

  zoo4:
    image: zookeeper:3.6.3
    restart: unless-stopped
    hostname: zoo4
    container_name: zoo4
    ports:
      - 2185:2181
    environment:
      ZOO_MY_ID: 4
      PEER_TYPE: observer  
      ZOO_SERVERS: server.1=zoo1:2888:3888;2181 server.2=zoo2:2888:3888;2181 server.3=zoo3:2888:3888;2181 server.4=0.0.0.0:2888:3888:observer;2181
    volumes:
      - ./zoo4/data:/data
      - ./zoo4/datalog:/datalog
    networks:
      - docker_net

创建网络：

docker network create docker_net

启动集群：

docker-compose -f docker-compose-zookeeper-cluster.yml up -d

docker exec -it zoo1 /bin/sh进入指定的容器执行 bash，使用zkServer.sh status 命令查看每个实例的服务状态：

$ docker exec -it zoo1 /bin/sh
# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /conf/zoo.cfg
Client port found: 2181. Client address: localhost. Client SSL: false.
Mode: follower

$ docker exec -it zoo2 /bin/sh
# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /conf/zoo.cfg
Client port found: 2181. Client address: localhost. Client SSL: false.
Mode: follower

$ docker exec -it zoo3 /bin/sh
# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /conf/zoo.cfg
Client port found: 2181. Client address: localhost. Client SSL: false.
Mode: leader

$ docker exec -it zoo4 /bin/sh
# zkServer.sh status
ZooKeeper JMX enabled by default
Using config: /conf/zoo.cfg
Client port found: 2181. Client address: localhost. Client SSL: false.
Mode: observer

zoo1、zoo2为 Follower，zoo3为 Leader，zoo4为 Observer。

使用 curator 连接集群并创建数据：

public class CuratorDemo1 {
    public static void main(String[] args) throws Exception {
        String zookeeperConnectionString = "192.168.3.6:2182,192.168.3.6:2183,192.168.3.6:2184";
        RetryPolicy retryPolicy = new ExponentialBackoffRetry(1000, 3);
        CuratorFramework curatorFramework = CuratorFrameworkFactory
                .newClient(zookeeperConnectionString, retryPolicy);
        curatorFramework.start();

        curatorFramework.create().forPath("/curator-node-01", "hello-world".getBytes());
                System.in.read();
    }
}

Zookeeper 集群模式