|
作者：Grey
原文地址: ZooKeeper学习笔记四：使用ZooKeeper实现一个简单的分布式锁
前置知识
完成ZooKeeper集群搭建以及熟悉ZooKeeperAPI基本使用
需求
当多个进程不在同一个系统中，用分布式锁控制多个进程对资源的访问。
在单机情况下，可以使用JUC包里面的工具来进行互斥控制。
但是在分布式系统后，由于分布式系统多线程、多进程并且分布在不同机器上，这将使原单机并发控制锁策略失效，为了解决这个问题就需要一种跨JVM的互斥机制来控制共享资源的访问，这就是分布式锁的由来。
当多个进程不在同一个系统中，就需要用分布式锁控制多个进程对资源的访问。
我们可以用ZooKeeper来模拟实现一个简单的分布式锁
环境准备
一个zk集权，ip和端口分别为：
- 192.168.205.145:2181
- 192.168.205.146:2181
- 192.168.205.147:2181
- 192.168.205.148:2181
定义主方法
App.java
public class App {
public static void main(String[] args) {
for (int i = 0; i < 10; i++) {
new Thread(() -> {
ZkLock lock = new ZkLock();
lock.lock();
System.out.println(Thread.currentThread().getName() + " doing work");
lock.release();
}).start();
}
while (true) {
}
}
}
如上，我们设计了一个ZkLock，其中lock方法是锁定资源，release方法是释放资源，我们并发了10个线程并发访问来模拟。
public class ZkLock implements AsyncCallback.StringCallback, Watcher, AsyncCallback.StatCallback, AsyncCallback.Children2Callback {
private CountDownLatch latch;
private ZooKeeper zk;
private String identify;
private String lockPath;
private String pathName;
public ZkLock() {
identify = Thread.currentThread().getName();
lockPath = "/lock";
latch = new CountDownLatch(1);
zk = ZookeeperConfig.create(ADDRESS + "/testLock");
}
public void lock() {
try {
zk.create(lockPath, currentThread().getName().getBytes(UTF_8), OPEN_ACL_UNSAFE, EPHEMERAL_SEQUENTIAL, this, currentThread().getName());
latch.await();
} catch (InterruptedException e) {
e.printStackTrace();
}
}
public void release() {
try {
zk.delete(pathName, -1);
System.out.println(identify + " over work....");
} catch (InterruptedException | KeeperException e) {
e.printStackTrace();
}
}
@Override
public void processResult(int rc, String path, Object ctx, String name) {
if (null != name) {
System.out.println(identify + " created " + name);
pathName = name;
zk.getChildren("/", false, this, "dasdfas");
}
}
@Override
public void processResult(int rc, String path, Object ctx, List<String> children, Stat stat) {
sort(children);
int i = children.indexOf(pathName.substring(1));
if (i == 0) {
System.out.println(identify + " first...");
try {
zk.setData("/", identify.getBytes(UTF_8), -1);
} catch (KeeperException | InterruptedException e) {
e.printStackTrace();
}
latch.countDown();
} else {
zk.exists("/" + children.get(i - 1), this, this, "ddsdf");
}
}
@Override
public void process(WatchedEvent event) {
switch (event.getType()) {
case None:
break;
case NodeCreated:
break;
case NodeDeleted:
zk.getChildren("/", false, this, "sdf");
break;
case NodeDataChanged:
break;
case NodeChildrenChanged:
break;
}
}
@Override
public void processResult(int rc, String path, Object ctx, Stat stat) {
}
}
关于上述代码的说明，我们规定创建的zk目录为/testLock,所以我们可以通过zk客户端在集群中先把/testLock目录建好，后续线程争抢的时候，我们只需要创建序列化的临时节点（以/lock开头），因为是序列化的，所以我们可以设置让第一个创建好节点的线程抢到锁，其他的线程排队等待。
所以lock方法实现如下：
zk.create(lockPath, currentThread().getName().getBytes(UTF_8), OPEN_ACL_UNSAFE, EPHEMERAL_SEQUENTIAL, this, currentThread().getName());
lock方法在执行的时候，会有一个回调，即：当节点创建成功后，会判断/testLock节点中有没有已经创建好的且在当前节点之前的节点，有的话，则注册一个一个对于/testLock目录的监听：
@Override
public void processResult(int rc, String path, Object ctx, String name) {
if (null != name) {
System.out.println(identify + " created " + name);
pathName = name;
zk.getChildren("/", false, this, "dasdfas");
}
}
一旦发现/testLock目录下已经有节点了，那么我们拿到/testLock下的所有节点，并排序，取最小的那个节点执行即可：
@Override
public void processResult(int rc, String path, Object ctx, List<String> children, Stat stat) {
sort(children);
int i = children.indexOf(pathName.substring(1));
if (i == 0) {
System.out.println(identify + " first...");
try {
zk.setData("/", identify.getBytes(UTF_8), -1);
} catch (KeeperException | InterruptedException e) {
e.printStackTrace();
}
latch.countDown();
} else {
zk.exists("/" + children.get(i - 1), this, this, "ddsdf");
}
}
release方法很简单，只需要把当前执行完毕的节点删除即可：
public void release() {
try {
zk.delete(pathName, -1);
System.out.println(identify + " over work....");
} catch (InterruptedException | KeeperException e) {
e.printStackTrace();
}
}
运行效果
确保zk中有/testLock这个节点，如果没有，请先创建一个:
Run App.java
可以看到控制台输出：
Thread-5 created /lock0000000000
Thread-4 created /lock0000000001
Thread-1 created /lock0000000002
Thread-9 created /lock0000000003
Thread-6 created /lock0000000004
Thread-2 created /lock0000000005
Thread-3 created /lock0000000006
Thread-0 created /lock0000000007
Thread-8 created /lock0000000008
Thread-7 created /lock0000000009
Thread-5 first...
Thread-5 doing work
Thread-5 over work....
Thread-4 first...
Thread-4 doing work
Thread-4 over work....
Thread-1 first...
Thread-1 doing work
Thread-1 over work....
Thread-9 first...
Thread-9 doing work
Thread-9 over work....
Thread-6 first...
Thread-6 doing work
Thread-6 over work....
Thread-2 first...
Thread-2 doing work
Thread-2 over work....
Thread-3 first...
Thread-3 doing work
Thread-3 over work....
Thread-0 first...
Thread-0 doing work
Thread-0 over work....
Thread-8 first...
Thread-8 doing work
Thread-8 over work....
Thread-7 first...
Thread-7 doing work
Thread-7 over work....
完整代码
Github
|