Mastering ExecutorService Shutdown: Tracking ThreadPool Termination

Let's say you want to execute some tasks. Since executing it through a single thread might take you quite some time to get the result, you decide to use the ever dependable ExecutorService to process it through multiple threads.

Here's a sample:

public static void main(String[] args) {
    ExecutorService executorService = Executors.newFixedThreadPool(5);
    for (int i = 0; i < 5; i++) {
        int temp = i;
        executorService.submit(() -> {
            task(temp);
        });
    }
    executorService.shutdown();
    System.out.println("ExecutorService is shutdown");
}

private static void task(int temp) {
    try {
        TimeUnit.SECONDS.sleep(1L);
        System.out.println("Task " + temp + " completed");
    } catch (InterruptedException e) {
        throw new RuntimeException(e);
    }
}

Of course, as usual, no example of Threads is ever complete without using "sleep" as an archetype of task execution.

It outputs,

ExecutorService is shutdown
Task 1 completed
Task 2 completed
Task 0 completed
Task 4 completed
Task 3 completed

Now imagine there's an endless queue of tasks, the number of which you don't know about. Maybe they are determined by the number of entries in a database which get added dynamically.

For example, a bank, wherein which it has to process a number of transactions throughout the day. The transaction end time will be 5 PM, beyond which it will not accept any additional tasks.

However, you do know that the number of tasks will be finite, and will end at some point of time.

How do you know the point in time when all the tasks have completed?

If you notice the above code snippet, the ExecutorService.shutdown() enables the main thread to exit immediately, but the background threads still process the accepted tasks to completion. Is there a way when you can get notified about the completion?

A couple of solutions come to mind:

1. Use a CountDownLatch to count the tasks - but since you don't know the number of tasks, it's impractical to use it.
2. Use ExecutorService.awaitTermination(). However, the time here is still undeterministic. You can use a very liberal ExecutorService.awaitTermination(Long.MAX_VALUE, TimeUnit.DAYS) or something similar. But that is again a blocking call.

Is there a better way to solve this?

Java does provide a better and relatively unknown way to get around this. The "trick" here is to know that Executors.newFixedThreadPool is essentially a ThreadPoolExecutor with predefined values. Let's check the implementation of Executors.newFixedThreadPool.

public static ExecutorService newFixedThreadPool(int nThreads) {
    return new ThreadPoolExecutor(nThreads, nThreads,
                                  0L, TimeUnit.MILLISECONDS,
                                  new LinkedBlockingQueue<Runnable>());
}

I really recommend to read the doc for ThreadPoolExecutor here. ExecutorService is a convenient wrapper over ThreadPoolExecutor.

... programmers are urged to use the more convenient Executors factory methods Executors.newCachedThreadPool() (unbounded thread pool, with automatic thread reclamation), Executors.newFixedThreadPool(int) (fixed size thread pool) and Executors.newSingleThreadExecutor() (single background thread), that preconfigure settings for the most common usage scenarios.

The section that will help us resolve out problem is:

Hook methods

This class provides protected overridable beforeExecute(Thread, Runnable) and afterExecute(Runnable, Throwable) methods that are called before and after execution of each task. These can be used to manipulate the execution environment; for example, reinitializing ThreadLocals, gathering statistics, or adding log entries. Additionally, method terminated() can be overridden to perform any special processing that needs to be done once the Executor has fully terminated.

We can use the terminated method to notify us of the same! But how do we use it?

public static void main(String[] args) {
    ExecutorService executorService = new ThreadPoolExecutor(5, 5,
            0L, TimeUnit.MILLISECONDS,
            new LinkedBlockingQueue<>()) {
        @Override
        protected void terminated() {
            super.terminated();
            System.out.println("ExecutorService is terminated");
        }
    };
    for (int i = 0; i < 5; i++) {
        int temp = i;
        executorService.submit(() -> {
            task(temp);
        });
    }
    executorService.shutdown();
    System.out.println("ExecutorService is shutdown");
}

private static void task(int temp) {
    try {
        TimeUnit.SECONDS.sleep(1L);
        System.out.println("Task " + temp + " completed");
    } catch (InterruptedException e) {
        throw new RuntimeException(e);
    }
}

If you do not prefer Anonymous classes (like me), you can always extend ThreadPoolExecutor yourself to create a custom one.

public static void main(String[] args) {
    ExecutorService executorService = getThreadPoolExecutor();
    for (int i = 0; i < 5; i++) {
        int temp = i;
        executorService.submit(() -> {
            task(temp);
        });
    }
    executorService.shutdown();
    System.out.println("ExecutorService is shutdown");
}

private static ThreadPoolExecutor getThreadPoolExecutor() {
    return new CustomThreadPoolExecutor(5, 5,
            0L, TimeUnit.MILLISECONDS,
            new LinkedBlockingQueue<>());
}

static class CustomThreadPoolExecutor extends ThreadPoolExecutor {
    public CustomThreadPoolExecutor(int corePoolSize, int maximumPoolSize, long keepAliveTime, TimeUnit unit, BlockingQueue<Runnable> workQueue) {
        super(corePoolSize, maximumPoolSize, keepAliveTime, unit, workQueue);
    }

    @Override
    protected void terminated() {
        super.terminated();
        System.out.println("ExecutorService is terminated");
    }
}

private static void task(int temp) {
    try {
        TimeUnit.SECONDS.sleep(1L);
        System.out.println("Task " + temp + " completed");
    } catch (InterruptedException e) {
        throw new RuntimeException(e);
    }
}

Here's the output to verify if it works as per our expectations.

ExecutorService is shutdown
Task 0 completed
Task 3 completed
Task 2 completed
Task 4 completed
Task 1 completed
ExecutorService is terminated

What are some other relatively unknown snippets you use? Let me know in the comments!