标签:Zookeeper3.8 ,Curator5.5;
一、简介
ZooKeeper是一个集中的服务,用于维护配置信息、命名、提供分布式同步、提供组服务。分布式应用程序以某种形式使用所有这些类型的服务。
二、环境搭建
1、修改配置文件
# 1、拷贝一份样本配置文件
cp zookeeper-3.8.3/conf/zoo_sample.cfg zookeeper-3.8.3/conf/zoo.cfg
# 2、修改数据文件地址,注意这里用本地路径
dataDir=/local-path/zookeeper-3.8.3/data
# 3、添加一个配置,处理启动日志的提示:ZooKeeper audit is disabled.
audit.enable=true
2、服务启动
# 1、启动服务端
zookeeper-3.8.3/bin/zkServer.sh start
# 2、停止服务端
zookeeper-3.8.3/bin/zkServer.sh stop
# 3、启动客户端
zookeeper-3.8.3/bin/zkCli.sh
3、客户端测几个增删查的命令
[zk: localhost:2181(CONNECTED) 0] create /cicada smile1
Created /cicada
[zk: localhost:2181(CONNECTED) 1] get /cicada
smile1
[zk: localhost:2181(CONNECTED) 2] ls /
[cicada, zookeeper]
[zk: localhost:2181(CONNECTED) 3] delete /cicada
三、工程搭建
1、工程结构
2、依赖管理
Curator是一组Java库,它让ZooKeeper的使用变得更加容易,这里的依赖实际是查询匹配版本的时候走了个捷径,也可以参考integration-redis
包,熟悉下Spring的封装策略。
<dependency>
<groupId>org.apache.zookeepergroupId>
<artifactId>zookeeperartifactId>
<version>${zookeeper.version}version>
dependency>
<dependency>
<groupId>org.springframework.integrationgroupId>
<artifactId>spring-integration-zookeeperartifactId>
<version>${spring-integration.version}version>
dependency>
3、配置文件
配置脚本
zookeeper:
#服务器地址
connectString: 127.0.0.1:2181
#会话超时时间
sessionTimeoutMs: 3000
#连接超时时间
connectionTimeoutMs: 60000
#最大重试次数
maxRetries: 3
#初始休眠时间
baseSleepTimeMs: 1000
配置类
@Configuration
public class ZookeeperConfig {
@Value("${zookeeper.connectString}")
private String connectString;
@Value("${zookeeper.baseSleepTimeMs}")
private int baseSleepTimeMs;
@Value("${zookeeper.maxRetries}")
private int maxRetries ;
@Value("${zookeeper.connectionTimeoutMs}")
int connectionTimeoutMs ;
@Value("${zookeeper.sessionTimeoutMs}")
int sessionTimeoutMs ;
private static CuratorFramework client = null ;
/**
* 初始化
*/
@PostConstruct
public void init (){
// 重试策略
RetryPolicy policy = new ExponentialBackoffRetry(baseSleepTimeMs, maxRetries);
// 创建Curator
client = CuratorFrameworkFactory.builder()
.connectString(connectString)
.connectionTimeoutMs(connectionTimeoutMs)
.sessionTimeoutMs(sessionTimeoutMs)
.retryPolicy(policy).build();
//开启连接
client.start();
}
@Bean
public CuratorFramework getClient (){
return client ;
}
}
四、ZooKeeper用法
测试几个API方法,节点创建和添加数据,以及判断和查询数据,还有就是基于ZooKeeper提供的读写锁能力。
public class ConfigTest {
@Autowired
private CuratorFramework client ;
@Test
public void testCreate () throws Exception {
// 创建一个持久化节点,断开连接时不会自动删除
client.create().creatingParentsIfNeeded().withMode(CreateMode.PERSISTENT).forPath("/path1");
}
@Test
public void testExists () throws Exception {
// 判断节点是否存在,path2不存在所以stat2是null
Stat stat1 = client.checkExists().forPath("/path1");
System.out.println(stat1);
Stat stat2 = client.checkExists().forPath("/path2");
System.out.println(stat2);
}
@Test
public void testSetData () throws Exception {
// 设置节点数据
client.setData().forPath("/path1", "data1".getBytes(StandardCharsets.UTF_8));
}
@Test
public void testCreateAndSet () throws Exception {
// 创建一个持久化节点并设置节点数据
client.create().creatingParentsIfNeeded().withMode(CreateMode.PERSISTENT)
.forPath("/path3","data3".getBytes(StandardCharsets.UTF_8));
}
@Test
public void testGetData () throws Exception {
// 查询节点数据
byte[] data = client.getData().forPath("/path3");
System.out.println(new String(data,StandardCharsets.UTF_8));
}
@Test
public void testDelete () throws Exception {
// 删除节点
client.delete().guaranteed().deletingChildrenIfNeeded().forPath("/path3");
}
@Test
public void testReadLock () throws Exception {
// 读写锁-读
InterProcessReadWriteLock lock = new InterProcessReadWriteLock(client,"/lock-read");
lock.readLock().acquire();
System.out.println("获取-ReadLock");
lock.readLock().release();
}
@Test
public void testWriteLock () throws Exception {
// 读写锁-写
InterProcessReadWriteLock lock = new InterProcessReadWriteLock(client,"/lock-write");
lock.writeLock().acquire();
System.out.println("获取-WriteLock");
lock.writeLock().release();
}
}
五、参考源码
文档仓库:
https://gitee.com/cicadasmile/butte-java-note
源码仓库:
https://gitee.com/cicadasmile/butte-spring-parent