Java 调度执行器线程进入等待状态

Question

因此，我尝试使用两个自定义线程和一个预定的线程执行器来实现从 Java 客户端到 Go 服务器的基本异步消息传递协议(protocol)。因此，我从 Java 客户端向 Go 服务器发起单个 TCP 套接字，并在两侧设置读取和写入缓冲区。

Writer线程将每个请求对应的唯一ID保存到Redis服务器中，并在TCP套接字上使用PrintWriter写入请求。

作家线程

public class WithdrawRequestWriter extends Thread {

    private static final int START_ACCOUNT_NO = 1;
    private static final int END_ACCOUNT_NO = 2_000;
    private static final int AMOUNT_TO_WITHDRAW = 100;

    private PrintWriter writer;
    private Jedis jedis;

    public WithdrawRequestWriter(PrintWriter writer, Jedis jedis) {
        this.writer = writer;
        this.jedis = jedis;
    }

    @Override
    public void run() {
        int i = 0;
        for (int accountNo = START_ACCOUNT_NO; accountNo <= END_ACCOUNT_NO; accountNo++) {
            String uniqueId = UUID.randomUUID().toString();
            if (jedis.set(uniqueId, String.valueOf(accountNo), "NX", "EX", 10) != null) {
                writer.write(uniqueId + " " + accountNo + " " + AMOUNT_TO_WITHDRAW + "\n");
                writer.flush();
            } else {
                System.err.println("Setting of UUID in redis failed: " + uniqueId);
            }
        }
    }
}

同样，读取器线程等待服务器的响应。一旦收到，它就会与 Redis 中的唯一 ID 匹配，将它们标记为完成(成功/失败)。

读者话题

public class WithdrawRequestReader extends Thread {

    private BufferedReader reader;
    private Jedis jedis;


    public WithdrawRequestReader(BufferedReader reader, Jedis jedis) {
        this.reader = reader;
        this.jedis = jedis;
    }

    @Override
    public void run() {
        String response;
        while (true) {
            try {
                if ((response = reader.readLine()) != null) {
                    String[] tokenizedResponse = response.split(" ");
                    System.out.println("Response is: " + response);
                    String value = jedis.get(tokenizedResponse[0]);
                    if (value != null) {
                        System.out.println("Request Token: " + tokenizedResponse[0]
                                + " | Status: " + tokenizedResponse[1]
                                + " | Account: " + tokenizedResponse[2]
                                + " | Amount: " + tokenizedResponse[3]
                                + " | Remaining Balance: " + tokenizedResponse[4]
                                + " | ResponeStr: " + tokenizedResponse[5]);
                    }
                } else {
                    System.out.println("Waiting for response...");
                }
            } catch (IOException ex) {
                ex.printStackTrace();
                System.exit(3);
            }
        }
    }
}

主线程

public class TestClient {

    private static final String HOSTNAME = "localhost";
    private static final int PORT = 12410;

    public static void main(String[] args) throws IOException {

        Socket socket = new Socket(HOSTNAME, PORT);
        PrintWriter writer = new PrintWriter(socket.getOutputStream());
        BufferedReader reader = new BufferedReader(new InputStreamReader(socket.getInputStream()));
        Jedis jedis = new Jedis("localhost");

        ScheduledExecutorService scheduledExecutorService = Executors.newScheduledThreadPool(2);
//        ScheduledExecutorService scheduledExecutorService = Executors.newSingleThreadScheduledExecutor();
        WithdrawRequestReader readerThread = new WithdrawRequestReader(reader, jedis);
        WithdrawRequestWriter writerThread = new WithdrawRequestWriter(writer, jedis);
        scheduledExecutorService.schedule(writerThread, 0, TimeUnit.MILLISECONDS);
        scheduledExecutorService.schedule(readerThread, 0, TimeUnit.MILLISECONDS);
    }
}

当我使用 singlePoolThreadExecutor 时，任务按顺序执行(如预期)，一切都很顺利。但是，当我将 threadPoolExecutor 与两个线程一起使用时，经过几次读写后，它们无限期地进入等待状态。我通过在线可用的免费工具检查了是否存在僵局，但没有得到任何帮助。下面是当它们被卡住时的线程转储。

线程转储

Full thread dump Java HotSpot(TM) 64-Bit Server VM (9.0.4+11 mixed mode):

"DestroyJavaVM" #16 prio=5 os_prio=31 tid=0x00007f974e06d000 nid=0x2503 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"pool-1-thread-2" #15 prio=5 os_prio=31 tid=0x00007f974e06c000 nid=0x6003 waiting on condition [0x000070000c819000]
   java.lang.Thread.State: WAITING (parking)
    at jdk.internal.misc.Unsafe.park(java.base@9.0.4/Native Method)
    - parking to wait for  <0x0000000747cf64b0> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
    at java.util.concurrent.locks.LockSupport.park(java.base@9.0.4/LockSupport.java:194)
    at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(java.base@9.0.4/AbstractQueuedSynchronizer.java:2062)
    at java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take(java.base@9.0.4/ScheduledThreadPoolExecutor.java:1119)
    at java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take(java.base@9.0.4/ScheduledThreadPoolExecutor.java:848)
    at java.util.concurrent.ThreadPoolExecutor.getTask(java.base@9.0.4/ThreadPoolExecutor.java:1092)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(java.base@9.0.4/ThreadPoolExecutor.java:1152)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(java.base@9.0.4/ThreadPoolExecutor.java:641)
    at java.lang.Thread.run(java.base@9.0.4/Thread.java:844)

"pool-1-thread-1" #14 prio=5 os_prio=31 tid=0x00007f974e8a9800 nid=0x5e03 waiting on condition [0x000070000c716000]
   java.lang.Thread.State: WAITING (parking)
    at jdk.internal.misc.Unsafe.park(java.base@9.0.4/Native Method)
    - parking to wait for  <0x0000000747cf64b0> (a java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject)
    at java.util.concurrent.locks.LockSupport.park(java.base@9.0.4/LockSupport.java:194)
    at java.util.concurrent.locks.AbstractQueuedSynchronizer$ConditionObject.await(java.base@9.0.4/AbstractQueuedSynchronizer.java:2062)
    at java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take(java.base@9.0.4/ScheduledThreadPoolExecutor.java:1119)
    at java.util.concurrent.ScheduledThreadPoolExecutor$DelayedWorkQueue.take(java.base@9.0.4/ScheduledThreadPoolExecutor.java:848)
    at java.util.concurrent.ThreadPoolExecutor.getTask(java.base@9.0.4/ThreadPoolExecutor.java:1092)
    at java.util.concurrent.ThreadPoolExecutor.runWorker(java.base@9.0.4/ThreadPoolExecutor.java:1152)
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(java.base@9.0.4/ThreadPoolExecutor.java:641)
    at java.lang.Thread.run(java.base@9.0.4/Thread.java:844)

"Service Thread" #11 daemon prio=9 os_prio=31 tid=0x00007f974d827000 nid=0xa003 runnable [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"Monitor Ctrl-Break" #10 daemon prio=5 os_prio=31 tid=0x00007f974e078800 nid=0x5903 runnable [0x000070000c40d000]
   java.lang.Thread.State: RUNNABLE
    at java.net.SocketInputStream.socketRead0(java.base@9.0.4/Native Method)
    at java.net.SocketInputStream.socketRead(java.base@9.0.4/SocketInputStream.java:116)
    at java.net.SocketInputStream.read(java.base@9.0.4/SocketInputStream.java:171)
    at java.net.SocketInputStream.read(java.base@9.0.4/SocketInputStream.java:141)
    at sun.nio.cs.StreamDecoder.readBytes(java.base@9.0.4/StreamDecoder.java:284)
    at sun.nio.cs.StreamDecoder.implRead(java.base@9.0.4/StreamDecoder.java:326)
    at sun.nio.cs.StreamDecoder.read(java.base@9.0.4/StreamDecoder.java:178)
    - locked <0x0000000747d9ea78> (a java.io.InputStreamReader)
    at java.io.InputStreamReader.read(java.base@9.0.4/InputStreamReader.java:185)
    at java.io.BufferedReader.fill(java.base@9.0.4/BufferedReader.java:161)
    at java.io.BufferedReader.readLine(java.base@9.0.4/BufferedReader.java:326)
    - locked <0x0000000747d9ea78> (a java.io.InputStreamReader)
    at java.io.BufferedReader.readLine(java.base@9.0.4/BufferedReader.java:392)
    at com.intellij.rt.execution.application.AppMainV2$1.run(AppMainV2.java:64)

"Common-Cleaner" #9 daemon prio=8 os_prio=31 tid=0x00007f974e00a000 nid=0xa203 in Object.wait() [0x000070000c30a000]
   java.lang.Thread.State: TIMED_WAITING (on object monitor)
    at java.lang.Object.wait(java.base@9.0.4/Native Method)
    - waiting on <0x0000000747f37610> (a java.lang.ref.ReferenceQueue$Lock)
    at java.lang.ref.ReferenceQueue.remove(java.base@9.0.4/ReferenceQueue.java:151)
    - waiting to re-lock in wait() <0x0000000747f37610> (a java.lang.ref.ReferenceQueue$Lock)
    at jdk.internal.ref.CleanerImpl.run(java.base@9.0.4/CleanerImpl.java:148)
    at java.lang.Thread.run(java.base@9.0.4/Thread.java:844)
    at jdk.internal.misc.InnocuousThread.run(java.base@9.0.4/InnocuousThread.java:122)

"Sweeper thread" #8 daemon prio=9 os_prio=31 tid=0x00007f974d81c000 nid=0x5803 runnable [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"C1 CompilerThread2" #7 daemon prio=9 os_prio=31 tid=0x00007f974d81b000 nid=0xa503 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE
   No compile task

"C2 CompilerThread1" #6 daemon prio=9 os_prio=31 tid=0x00007f974e005000 nid=0xa603 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE
   No compile task

"C2 CompilerThread0" #5 daemon prio=9 os_prio=31 tid=0x00007f974d81a800 nid=0xa803 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE
   No compile task

"Signal Dispatcher" #4 daemon prio=9 os_prio=31 tid=0x00007f974e01a800 nid=0x5603 waiting on condition [0x0000000000000000]
   java.lang.Thread.State: RUNNABLE

"Finalizer" #3 daemon prio=8 os_prio=31 tid=0x00007f974d041800 nid=0x4003 in Object.wait() [0x000070000bc75000]
   java.lang.Thread.State: WAITING (on object monitor)
    at java.lang.Object.wait(java.base@9.0.4/Native Method)
    - waiting on <0x0000000747f0d078> (a java.lang.ref.ReferenceQueue$Lock)
    at java.lang.ref.ReferenceQueue.remove(java.base@9.0.4/ReferenceQueue.java:151)
    - waiting to re-lock in wait() <0x0000000747f0d078> (a java.lang.ref.ReferenceQueue$Lock)
    at java.lang.ref.ReferenceQueue.remove(java.base@9.0.4/ReferenceQueue.java:172)
    at java.lang.ref.Finalizer$FinalizerThread.run(java.base@9.0.4/Finalizer.java:216)

"Reference Handler" #2 daemon prio=10 os_prio=31 tid=0x00007f974d040800 nid=0x4103 waiting on condition [0x000070000bb72000]
   java.lang.Thread.State: RUNNABLE
    at java.lang.ref.Reference.waitForReferencePendingList(java.base@9.0.4/Native Method)
    at java.lang.ref.Reference.processPendingReferences(java.base@9.0.4/Reference.java:174)
    at java.lang.ref.Reference.access$000(java.base@9.0.4/Reference.java:44)
    at java.lang.ref.Reference$ReferenceHandler.run(java.base@9.0.4/Reference.java:138)

"VM Thread" os_prio=31 tid=0x00007f974e884000 nid=0x4303 runnable 

"GC Thread#0" os_prio=31 tid=0x00007f974e80d800 nid=0x2a03 runnable 

"GC Thread#1" os_prio=31 tid=0x00007f974e80e800 nid=0x5303 runnable 

"GC Thread#2" os_prio=31 tid=0x00007f974d80c800 nid=0x2b03 runnable 

"GC Thread#3" os_prio=31 tid=0x00007f974d000800 nid=0x2c03 runnable 

"G1 Main Marker" os_prio=31 tid=0x00007f974e835800 nid=0x4803 runnable 

"G1 Marker#0" os_prio=31 tid=0x00007f974d80d800 nid=0x4503 runnable 

"G1 Refine#0" os_prio=31 tid=0x00007f974e80f800 nid=0x4a03 runnable 

"G1 Refine#1" os_prio=31 tid=0x00007f974d80d000 nid=0x4c03 runnable 

"G1 Refine#2" os_prio=31 tid=0x00007f974e80f000 nid=0x4e03 runnable 

"G1 Refine#3" os_prio=31 tid=0x00007f974e806800 nid=0x5003 runnable 

"G1 Young RemSet Sampling" os_prio=31 tid=0x00007f974e810800 nid=0x2f03 runnable 

"VM Periodic Task Thread" os_prio=31 tid=0x00007f974e079800 nid=0x5c03 waiting on condition 

JNI global references: 421

Heap
 garbage-first heap   total 131072K, used 10240K [0x0000000740000000, 0x0000000740100400, 0x00000007c0000000)
  region size 1024K, 11 young (11264K), 0 survivors (0K)
 Metaspace       used 8824K, capacity 9274K, committed 9472K, reserved 1058816K
  class space    used 832K, capacity 927K, committed 1024K, reserved 1048576K

我相信Go服务器不是问题，否则单线程执行器也应该卡住，但事实并非如此。可能我遇到了一些僵局，但无法找出确切的原因。有人可以帮忙吗？如果您还想要什么，请告诉我。

Answer 1

代码的套接字和流部分似乎是正确的。并行使用套接字的输入和输出流是正确的，因为它们是独立的流。

问题出在 Jedis 组件上。 Jedis 不是线程安全的。 Jedis 文档说:

You shouldn't use the same instance from different threads because you'll have strange errors. And sometimes creating lots of Jedis instances is not good enough because it means lots of sockets and connections, which leads to strange errors as well. A single Jedis instance is not threadsafe!...

需要重写代码，让每个线程都有一个私有(private)的 Jedis 实例。