为什么在 zookeeper WorkerSender 中使用 poll() 而不是 take()

Question

我一直在阅读 zookeeper 源代码，发现当从 FastLeaderElection.WorkerSender 中的队列中获取 ToSend 消息时，他们使用 poll() 而不是 take()。为什么？

poll()更好吗？

搜索 git 历史，我发现它开始使用 take()，所以必须有一个原因更改为 poll()。

public void run() {
  while (!stop) {
    try {
      ToSend m = sendqueue.poll(3000, TimeUnit.MILLISECONDS);
      if (m == null) {
        continue;
      }
      process(m);
    } catch (InterruptedException e) {
      break;
    }
  }
  LOG.info("WorkerSender is down");
}

Answer 1

TL;DR: 检查停止标志。

为 ZOOKEEPER-498: Diff of FastLeaderElection.java

所做的更改

旧代码：

class WorkerReceiver implements Runnable {

    QuorumCnxManager manager;

    WorkerReceiver(QuorumCnxManager manager) {
        this.manager = manager;
    }

    public void run() {

        Message response;
        while (true) {
            // Sleeps on receive
            try{
                response = manager.recvQueue.take();

                // Receive new message
                LOG.debug("Receive new message.");

新代码：

class WorkerReceiver implements Runnable {
    volatile boolean stop;
    QuorumCnxManager manager;

    WorkerReceiver(QuorumCnxManager manager) {
        this.stop = false;
        this.manager = manager;
    }

    public void run() {

        Message response;
        while (!stop) {
            // Sleeps on receive
            try{
                response = manager.recvQueue.poll(3000, TimeUnit.MILLISECONDS);
                if(response == null) continue;

                // Receive new message
                LOG.debug("Receive new message.");

they use poll() instead of take(). Why ?

因此他们可以每 3 秒检查一次 新 stop 标志。

为什么在 zookeeper WorkerSender 中使用 poll() 而不是 take()

Why use poll() instead of take() in zookeeper WorkerSender

java

apache-zookeeper