Apache Curator

Apache Curator 是Apache ZooKeeper 的Java 客户端库。Curator 项目的目标是简化ZooKeeper 客户端的使用。例如，在以前的代码展示中，我们都要自己处理
ConnectionLossException 。另外Curator 为常见的分布式协同服务提供了高质量的实现。Apache Curator 最初是Netflix 研发的，后来捐献给了Apache 基金会，目前是Apache 的顶级项目。

Curator 技术栈

Client：封装了 Zookeeper 类，管理和Zookeeper集群的连接，并提供了重建连接机制。

Framework：为所有的Zookeeper操作提供了重试机制，对提供了一个Fluent风格的API。

Recipes：使用framework实现了大量 Zookeeper 协同服务。

Extensions：扩展模块

Client 的创建，初始化一个client分为两步骤：

创建client
启动client

//创建 client
RetryPolicy retryPolicy = new ExponentialBackoffRetry(1000, 3);// 使用 Factory 方法
CuratorFramework client = CuratorFrameworkFactory.newClient(connectString, retryPolicy);
// Fluent 风格
CuratorFramework client = CuratorFrameworkFactory.builder().connectString(connectString).retryPolicy(retryPolicy).build();
//启动 client
client.start();

Fluent 风格API

// 同步版本
client.create().withMode(CreateMode.PERSISTENT).forPath(path, data);
// 异步版本
client.create().withMode(CreateMode.PERSISTENT).inBackground().forPath(path, data);
// 使用watch
client.getData().watched().forPath(path);

案例：

public class TestCurator {private CuratorFramework client = null;private String connectString = "192.168.1.100:2181";private RetryPolicy retryPolicy = new ExponentialBackoffRetry(1000, 3);@Beforepublic void creatClient() {client = CuratorFrameworkFactory.builder().connectString(connectString).retryPolicy(retryPolicy).build();client.start();}@Testpublic void testSyncOp() throws Exception {String path = "/one";byte[] data = "testSyncOp".getBytes();client.create().withMode(CreateMode.PERSISTENT).forPath(path,data);byte[] actualData = client.getData().forPath(path);System.out.println(new String(actualData));client.delete().forPath(path);client.close();}@Testpublic void testAsyncOp() throws Exception {String path = "/two";final byte[] data = "testAsyncOp".getBytes();final CountDownLatch latch = new CountDownLatch(1);client.getCuratorListenable().addListener((CuratorFramework c, CuratorEvent event) -> {switch (event.getType()) {case CREATE:System.out.printf("znode '%s' created\n", event.getPath());c.getData().inBackground().forPath(event.getPath());break;case GET_DATA:System.out.printf("got the data of znode '%s'\n", event.getPath());System.out.println(new String(event.getData()));c.delete().inBackground().forPath(path);break;case DELETE:System.out.printf("znode '%s' deleted.\n", event.getPath());latch.countDown();break;}});client.create().withMode(CreateMode.PERSISTENT).inBackground().forPath(path, data);latch.await();client.close();}@Testpublic void testWatch() throws Exception {String path = "/three";byte[] data = {'3'};byte[] newData = {'4'};CountDownLatch latch = new CountDownLatch(1);client.getCuratorListenable().addListener(((CuratorFramework c, CuratorEvent event ) -> {switch (event.getType()) {case WATCHED:WatchedEvent we = event.getWatchedEvent();System.out.println("watched event:" + we);if (we.getType() == Watcher.Event.EventType.NodeDataChanged&& we.getPath().equals(path)) {System.out.println("got the event for the triggered watch");byte[] actualData = c.getData().forPath(path);System.out.println(new String(actualData));}latch.countDown();break;}}));client.create().withMode(CreateMode.PERSISTENT).forPath(path, data);byte[] actualData = client.getData().watched().forPath(path);System.out.println(new String(actualData));client.setData().forPath(path, newData);latch.await();client.delete().forPath(path);}@Testpublic void testCallbackAndWatch() throws Exception {String path = "/four";byte[] data = {'4'};byte[] newData = {'5'};CountDownLatch latch = new CountDownLatch(3);client.getCuratorListenable().addListener(((CuratorFramework c, CuratorEvent event ) -> {switch (event.getType()) {case CREATE:System.out.printf("znode '%s' created\n", event.getPath());System.out.println(new String(client.getData().watched().forPath(path)));client.setData().forPath(path, newData);latch.countDown();break;case WATCHED:WatchedEvent we = event.getWatchedEvent();System.out.println("watched event:" + we);if (we.getType() == Watcher.Event.EventType.NodeDataChanged&& we.getPath().equals(path)) {System.out.println("got the event for the triggered watch");byte[] actualData = c.getData().forPath(path);System.out.println(new String(actualData));}latch.countDown();break;}}));client.create().withMode(CreateMode.PERSISTENT).inBackground().forPath(path, data);latch.await();client.delete().forPath(path);}
}

参考链接：https://time.geekbang.org/course/intro/100034201

Apache Curator相关推荐

Apache ZooKeeper - 使用Apache Curator操作ZK
文章目录原生ZK API VS Curator Curator 概述 Maven依赖会话创建静态工厂方式创建会话使用 fluent 风格创建会话创建节点 protection 模式 ,规避僵 ...
【zookeeper】Apache curator优点介绍
文章目录 1. 简介 2. 项目组件 2.1 版本 2.2 项目组件 2.3 Maven依赖 3. 案例及功能说明 3.1 创建会话 3.1.1 重试策略 3.1.2 创建节点 3.1.3 删除节点 ...
Apache Curator操作zookeeper的API使用
curator简介与客户端之间的异同点常用的zookeeper java客户端: zookeeper原生Java API zkclient Apache curator ZooKeeper原生Jav ...
基于Apache Curator框架的ZooKeeper基本用法详解
简介 Apache Curator是一个比较完善的ZooKeeper客户端框架,通过封装的一套高级API 简化了ZooKeeper的操作.通过查看官方文档,可以发现Curator主要解决了三类问题: ...
Apache Curator客户端的使用
当前已有的三种API客户端 zk原生API :不支持超时自动重连,不支持节点递归创建 zkclient:无文档,异常处理弱爆了(简单的抛出RuntimeException) Apache Curato ...
Apache Curator实战
Apache Curator入门实战 Curator是Netflix公司开源的一个Zookeeper客户端,与Zookeeper提供的原生客户端相比,Curator的抽象层次更高,简化了Zookeep ...
Dubbo 启动报java.lang.NoClassDefFoundError: org/apache/curator/RetryPolicy
1.## Dubbo 启动报java.lang.NoClassDefFoundError: org/apache/curator/RetryPolicy 2. Failed to subscribe ...
Caused by: java.lang.NoClassDefFoundError: org/apache/curator/framework/CuratorFrameworkFactory
springboot整合dubbo,apache的报错: 添加依赖: <dependency><groupId>org.apache.curator</groupId&g ...
Zookeeper 原生客户端、可视化工具 ZooInspector 、Apache Curator
目录 Zookeeper 原生客户端 Apache Curator 开源客户端可视化客户端工具 ZooInspector Zookeeper 原生客户端 1.类似 Redis 有多种 Java 客户 ...

Apache Curator

Curator 技术栈

Apache Curator相关推荐

最新文章

热门文章