Apache Curator Leader Election

阿新 • • 發佈：2019-01-18

用於Leader選舉，也可以用Shared Reentrant Lock來實現。

如果需要叢集中的固定的一臺機器去做的事，就可以用此特性來實現，直到這臺Leader死去，會產生新的Leader。

還有一種典型的場景，master-slave模式。也可以用Curator Leader Election輕鬆實現，包括充當maste的機器死掉後，會產生新的master，以及當新加入機器後，該機器會直接參與Leader競爭者。That's Awesome.

1.直接執行LLtest

package com.collonn.javaUtilMvn.zookeeper.curator.LeaderElection;

import com.collonn.javaUtilMvn.zookeeper.curator.NodeCache.NLClientCreate;
import com.collonn.javaUtilMvn.zookeeper.curator.NodeCache.NLClientDelete;
import com.collonn.javaUtilMvn.zookeeper.curator.NodeCache.NLClientUpdate;

public class LLTest {
    public static void main(String[] args) throws Exception {
        LeaderListener.main(null);
        LeaderListener2.main(null);
        LeaderListener3.main(null);
        LeaderListener4.main(null);
    }
}

package com.collonn.javaUtilMvn.zookeeper.curator.LeaderElection;

import org.apache.curator.RetryPolicy;
import org.apache.curator.framework.CuratorFramework;
import org.apache.curator.framework.CuratorFrameworkFactory;
import org.apache.curator.framework.recipes.cache.ChildData;
import org.apache.curator.framework.recipes.cache.PathChildrenCache;
import org.apache.curator.framework.recipes.cache.PathChildrenCacheEvent;
import org.apache.curator.framework.recipes.cache.PathChildrenCacheListener;
import org.apache.curator.framework.recipes.leader.LeaderSelector;
import org.apache.curator.framework.recipes.leader.LeaderSelectorListener;
import org.apache.curator.framework.state.ConnectionState;
import org.apache.curator.retry.ExponentialBackoffRetry;
import org.apache.curator.utils.EnsurePath;

import java.util.List;

public class LeaderListener {
    public static final String C_PATH = "/TestLeader";
    public static final String CHARSET = "UTF-8";
    public static final String APP_NAME = "app1";

    public static void main(String[] args) {
        try {
            new Thread(new Runnable() {
                @Override
                public void run() {
                    try{
                        String zookeeperConnectionString = "127.0.0.1:2181";
                        RetryPolicy retryPolicy = new ExponentialBackoffRetry(1000, 3);
                        CuratorFramework client = CuratorFrameworkFactory.newClient(zookeeperConnectionString, retryPolicy);
                        client.start();

                        //ensure path of /test
                        new EnsurePath(C_PATH).ensure(client.getZookeeperClient());

                        final LeaderSelector leaderSelector = new LeaderSelector(client, C_PATH, new LeaderSelectorListener() {
                            @Override
                            public void takeLeadership(CuratorFramework client) throws Exception {
                                try {
                                    int timeMilliSeconds = 6000;
                                    System.out.println("======" + APP_NAME + " take leader ship, will do some task, will hold time milli seconds=" + timeMilliSeconds);

                                    //once you take the leader ship
                                    //and you want hold the leader ship during the whole life of APP1
                                    //you should use Thread.sleep(Integer.MAX_VALUE)
                                    //once tha APP1 dead, the other APP (participants) will reElect an new leader
                                    for(int i = 0; i < 6; i++){
                                        System.out.println("===" + APP_NAME + " sleep " + i);
                                        Thread.sleep(1000);
                                    }
                                }catch (Exception e){
                                    e.printStackTrace();
                                }
                            }

                            @Override
                            public void stateChanged(CuratorFramework client, ConnectionState newState) {

                            }
                        });

                        leaderSelector.start();

                        Thread.sleep(Integer.MAX_VALUE);
                        client.close();
                    }catch (Exception e){
                        e.printStackTrace();
                    }
                }
            }).start();
        }catch (Exception e){
            e.printStackTrace();
        }
    }
}

package com.collonn.javaUtilMvn.zookeeper.curator.LeaderElection;

import org.apache.curator.RetryPolicy;
import org.apache.curator.framework.CuratorFramework;
import org.apache.curator.framework.CuratorFrameworkFactory;
import org.apache.curator.framework.recipes.leader.LeaderSelector;
import org.apache.curator.framework.recipes.leader.LeaderSelectorListener;
import org.apache.curator.framework.state.ConnectionState;
import org.apache.curator.retry.ExponentialBackoffRetry;
import org.apache.curator.utils.EnsurePath;

public class LeaderListener2 {
    public static final String C_PATH = "/TestLeader";
    public static final String CHARSET = "UTF-8";
    public static final String APP_NAME = "app2";

    public static void main(String[] args) {
        try {
            new Thread(new Runnable() {
                @Override
                public void run() {
                    try{
                        String zookeeperConnectionString = "127.0.0.1:2181";
                        RetryPolicy retryPolicy = new ExponentialBackoffRetry(1000, 3);
                        CuratorFramework client = CuratorFrameworkFactory.newClient(zookeeperConnectionString, retryPolicy);
                        client.start();

                        //ensure path of /test
                        new EnsurePath(C_PATH).ensure(client.getZookeeperClient());

                        final LeaderSelector leaderSelector = new LeaderSelector(client, C_PATH, new LeaderSelectorListener() {
                            @Override
                            public void takeLeadership(CuratorFramework client) throws Exception {
                                try {
                                    int timeMilliSeconds = 6000;
                                    System.out.println("======" + APP_NAME + " take leader ship, will do some task, will hold time milli seconds=" + timeMilliSeconds);

                                    for(int i = 0; i < 6; i++){
                                        System.out.println("===" + APP_NAME + " sleep " + i);
                                        Thread.sleep(1000);
                                    }
                                }catch (Exception e){
                                    e.printStackTrace();
                                }
                            }

                            @Override
                            public void stateChanged(CuratorFramework client, ConnectionState newState) {

                            }
                        });

                        leaderSelector.start();

                        Thread.sleep(Integer.MAX_VALUE);
                        client.close();
                    }catch (Exception e){
                        e.printStackTrace();
                    }
                }
            }).start();
        }catch (Exception e){
            e.printStackTrace();
        }
    }
}

package com.collonn.javaUtilMvn.zookeeper.curator.LeaderElection;

import org.apache.curator.RetryPolicy;
import org.apache.curator.framework.CuratorFramework;
import org.apache.curator.framework.CuratorFrameworkFactory;
import org.apache.curator.framework.recipes.leader.LeaderSelector;
import org.apache.curator.framework.recipes.leader.LeaderSelectorListener;
import org.apache.curator.framework.state.ConnectionState;
import org.apache.curator.retry.ExponentialBackoffRetry;
import org.apache.curator.utils.EnsurePath;

public class LeaderListener3 {
    public static final String C_PATH = "/TestLeader";
    public static final String CHARSET = "UTF-8";
    public static final String APP_NAME = "app3";

    public static void main(String[] args) {
        try {
            new Thread(new Runnable() {
                @Override
                public void run() {
                    try{
                        String zookeeperConnectionString = "127.0.0.1:2181";
                        RetryPolicy retryPolicy = new ExponentialBackoffRetry(1000, 3);
                        CuratorFramework client = CuratorFrameworkFactory.newClient(zookeeperConnectionString, retryPolicy);
                        client.start();

                        //ensure path of /test
                        new EnsurePath(C_PATH).ensure(client.getZookeeperClient());

                        final LeaderSelector leaderSelector = new LeaderSelector(client, C_PATH, new LeaderSelectorListener() {
                            @Override
                            public void takeLeadership(CuratorFramework client) throws Exception {
                                try {
                                    int timeMilliSeconds = 6000;
                                    System.out.println("======" + APP_NAME + " take leader ship, will do some task, will hold time milli seconds=" + timeMilliSeconds);

                                    for(int i = 0; i < 6; i++){
                                        System.out.println("===" + APP_NAME + " sleep " + i);
                                        Thread.sleep(1000);
                                    }
                                }catch (Exception e){
                                    e.printStackTrace();
                                }
                            }

                            @Override
                            public void stateChanged(CuratorFramework client, ConnectionState newState) {

                            }
                        });

                        leaderSelector.start();

                        Thread.sleep(Integer.MAX_VALUE);
                        client.close();
                    }catch (Exception e){
                        e.printStackTrace();
                    }
                }
            }).start();
        }catch (Exception e){
            e.printStackTrace();
        }
    }
}

package com.collonn.javaUtilMvn.zookeeper.curator.LeaderElection;

import org.apache.curator.RetryPolicy;
import org.apache.curator.framework.CuratorFramework;
import org.apache.curator.framework.CuratorFrameworkFactory;
import org.apache.curator.framework.recipes.leader.LeaderSelector;
import org.apache.curator.framework.recipes.leader.LeaderSelectorListener;
import org.apache.curator.framework.state.ConnectionState;
import org.apache.curator.retry.ExponentialBackoffRetry;
import org.apache.curator.utils.EnsurePath;

public class LeaderListener4 {
    public static final String C_PATH = "/TestLeader";
    public static final String CHARSET = "UTF-8";
    public static final String APP_NAME = "app4";

    public static void main(String[] args) {
        try {
            new Thread(new Runnable() {
                @Override
                public void run() {
                    try{
                        String zookeeperConnectionString = "127.0.0.1:2181";
                        RetryPolicy retryPolicy = new ExponentialBackoffRetry(1000, 3);
                        CuratorFramework client = CuratorFrameworkFactory.newClient(zookeeperConnectionString, retryPolicy);
                        client.start();

                        //ensure path of /test
                        new EnsurePath(C_PATH).ensure(client.getZookeeperClient());

                        final LeaderSelector leaderSelector = new LeaderSelector(client, C_PATH, new LeaderSelectorListener() {
                            @Override
                            public void takeLeadership(CuratorFramework client) throws Exception {
                                try {
                                    int timeMilliSeconds = 6000;
                                    System.out.println("======" + APP_NAME + " take leader ship, will do some task, will hold time milli seconds=" + timeMilliSeconds);

                                    for(int i = 0; i < 6; i++){
                                        System.out.println("===" + APP_NAME + " sleep " + i);
                                        Thread.sleep(1000);
                                    }
                                }catch (Exception e){
                                    e.printStackTrace();
                                }
                            }

                            @Override
                            public void stateChanged(CuratorFramework client, ConnectionState newState) {

                            }
                        });

                        leaderSelector.start();

                        Thread.sleep(Integer.MAX_VALUE);
                        client.close();
                    }catch (Exception e){
                        e.printStackTrace();
                    }
                }
            }).start();
        }catch (Exception e){
            e.printStackTrace();
        }
    }
}

Apache Curator Leader Election

用於Leader選舉，也可以用Shared Reentrant Lock來實現。如果需要叢集中的固定的一臺機器去做的事，就可以用此特性來實現，直到這臺Leader死去，會產生新的Leader。還有一種典型的場景，master-slave模式。也可以用Curator Le

Zookeeper 學習筆記之 Leader Election

通知客戶就會 lec 搶占式類型二次 lead per ZooKeeper四種節點類型： Persist Persist_Sequential Ephemeral Ephemeral_Sequential 在節點上可註冊的Watch，客戶端先得到通知再得到數據，

Apache Curator操作zookeeper的API使用

zookeeper 分布式集群 curator 中間件 curator簡介與客戶端之間的異同點常用的zookeeper java客戶端： zookeeper原生Java API zkclient Apache curator ZooKeeper原生Java API的不足之處：在

如何運用zookepper進行kafka Leader Election?

註冊 bsp roo 可能 zookepper 監聽 election 如果 etc 主要有兩種方法： (一)搶註Leader節點-----非公平模式 (二)先到先得，後者監聽前者-----公平模式 (一)搶註Leader節點-----非公平模式 1.創建Leader父節點

【zookeeper】Apache curator的使用及zk分布式鎖實現

sets finally tac -- ont zkcli 單節點基本操作新建上篇,本篇主要講Apache開源的curator的使用,有了curator,利用Java對zookeeper的操作變得極度便捷. 其實在學之前我也有個疑慮,我為啥要學curator,撇開漲薪

【zookeeper】Apache curator的使用及zk分散式鎖實現

接上篇,本篇主要講Apache開源的curator的使用,有了curator,利用Java對zookeeper的操作變得極度便捷. 其實在學之前我也有個疑慮,我為啥要學curator,撇開漲薪這些外在的東西,就單技術層面來講,學curator能幫我做些什麼?這就不得不從zookeeper說起,上

Apache Curator客戶端

一：Apache Curator簡介 1. Curator主要從以下幾個方面降低了zk使用的複雜性重試機制:提供可插拔的重試機制, 它將給捕獲所有可恢復的異常配置一個重試策略,並且內部也提供了幾種標準的重試策略(比如指數補償) 連線狀態監控: Curator初始化

java.lang.NoClassDefFoundError: org/apache/curator/RetryPolicy解決方法

今天整合es-job到公司的框架時，啟動時出現上述錯誤 java.lang.NoClassDefFoundError: org/apache/curator/RetryPolicy at storm.kafka.KafkaSpout.open(KafkaSpout.java:68) at backtype.

zookeeper Apache Curator

1 簡介 Curator是Netflix公司開源的一套Zookeeper客戶端框架。瞭解過Zookeeper原生API都會清楚其複雜度。Curator幫助我們在其基礎上進行封裝、實現一些開發細節，包括接連重連、反覆註冊Watcher和NodeExistsExcep

第7章 Apache Curator客戶端的使用

Apache Curator客戶端的使用 7-1 curator簡介與客戶端之間的異同點 7-2 搭建maven工程，建立curator與zkserver的連線 7-3 zk名稱空間以及建立節點 7-9 zk-watcher例項統一更新N臺節點的配置

Kafka引數unclean.leader.election.enable詳解

如何提高Kafka可靠性是一個可以長篇大論的主題。很多初學者會簡單的認為將客戶端引數acks設定為-1即可保證Kafka的可靠性，顯然這是很片面的觀點。就可靠性本身而言，它並不是一個可以用“是”或者“否”來衡量的一個指標，而一般是用幾個9來衡量。就引數方面而言，與Kafka可靠性相關的引數不止ack

Apache Curator簡單介紹

提供了一個抽象級別更高的API，來操作Zookeeper，類似Guava提供的很多工具，讓Java書寫起來更加方便。至於有沒有用，那就要看每個人自己的理解了。 1、依賴 <dependency> <groupId>org.a

storm:ConnectionStateManager-0 WARN org.apache.curator.framework.state.ConnectionStateManager

strom異常解決： [ConnectionStateManager-0] WARN org.apache.curator.framework.state.ConnectionStateManager - There are no ConnectionStateListe

使用Apache Curator管理ZooKeeper

Apache ZooKeeper是為了幫助解決複雜問題的軟體工具，它可以幫助使用者從複雜的實現中解救出來。然而，ZooKeeper只暴露了原語，這取決於使用者如何使用這些原語來解決應用程式中的協調問題。社群已經在ZooKeeper資料模型及其API之上開發了

Apache Curator操作zookeeper的API使用——watcher

curator在註冊watch事件上，提供了一個usingWatcher方法，使用這個方法註冊的watch事件和預設watch事件一樣，監聽只會觸發一次，監聽完畢後就會銷燬，也就是一次性的。而這個方法有兩種引數可選，一個是zk原生API的Watcher介面的實現類，另一個是Curator提供的Cur

深入學習Kafka：Leader Election

本文所講的Leader是指叢集中的Controller，而不是各個Partition的Leader。為什麼要有Leader？在Kafka早期版本，對於分割槽和副本的狀態的管理依賴於zookeeper的Watcher和佇列：每一個broker都會在

使用Apache Curator實現服務的註冊和發現

使用zookeeper可以實現服務的註冊和發現，而Curator是對zookeeper進行的一層封裝，自然也封裝了一套實現服務的註冊和發現，本文就介紹如何使用Curator實現服務的註冊和發現首先要安裝zookeeper，我這裡安裝的是：zookeeper-3.4.6

【第三方類庫】Apache Curator

Curator是Netflix公司開源的一個Zookeeper客戶端，與Zookeeper提供的原生客戶端相比，Curator的抽象層次更高，簡化了Zookeeper客戶端的開發量。包含一下幾個模

Apache Curator簡單使用（一）

轉載自： http://ifeve.com/zookeeper-curato-framework/ http://blog.csdn.net/dc_726/article/details/46475633 http

Apache curator-recipes程式碼範例

Apache curator-recipes程式碼例項 ? ? Apache curator-recipes元件提供了大量已經"生產化"(produced)的特性,極大的簡化了使用zk的複雜度. ? ? 1. Cache: 提供了對一個Node持續監聽,如果節點資料變更,

Apache Curator Leader Election

相關推薦