我正在研究并行计算的软件开发框架JavaSeis.org。我需要一个强大的机制来报告线程异常。在开发过程中,了解异常的来源具有很高的价值,因此我想在过度报告方面犯错误。我也希望能够在线程中处理Junit4测试。方法是否合理或有更好的方法吗?
import java.util.concurrent.Callable;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Future;
public class TestThreadFailure {
public static void main(String[] args) {
int size = 1;
ExecutorService exec = Executors.newFixedThreadPool(size);
ThreadFailTask worker = new ThreadFailTask();
Future<Integer> result = exec.submit(worker);
try {
Integer value = result.get();
System.out.println("Result: " + value);
} catch (Throwable t) {
System.out.println("Caught failure: " + t.toString());
exec.shutdownNow();
System.out.println("Stack Trace:");
t.printStackTrace();
return;
}
throw new RuntimeException("Did not catch failure !!");
}
public static class ThreadFailTask implements Callable<Integer> {
@Override
public Integer call() {
int nbuf = 65536;
double[][] buf = new double[nbuf][nbuf];
return new Integer((int) buf[0][0]);
}
}
}
答案 0 :(得分:18)
考虑在execute()
上调用submit()
而不是ExecutorService
。使用execute()
调用的Thread
会在失败时调用Thread.UncaughtExceptionHandler
。
只需在Thread.UncaughtExceptionHandler
上设置一个ThreadFactory
,然后在Threads
上使用execute()
而非{{1}调用您的工作}。
答案 1 :(得分:7)
在使用submit()
时,我不相信有一个标准的“钩子”可以解决这些异常。但是,如果你需要支持submit()
(这听起来合理,假设你使用Callable
),你总是可以包装Callables和Runnables:
ExecutorService executor = new ThreadPoolExecutor(1, 10, 60, TimeUnit.SECONDS, new LinkedBlockingDeque<Runnable>()) {
@Override
public <T> Future<T> submit(final Callable<T> task) {
Callable<T> wrappedTask = new Callable<T>() {
@Override
public T call() throws Exception {
try {
return task.call();
}
catch (Exception e) {
System.out.println("Oh boy, something broke!");
e.printStackTrace();
throw e;
}
}
};
return super.submit(wrappedTask);
}
};
当然,这种方法只有在您构建ExecutorService
的情况下才有效。此外,请记住覆盖所有三个submit()
变体。
答案 2 :(得分:3)
如此线程What is the difference between submit and execute method with ThreadPoolExecutor中所述,使用execute只有在实现Runnable而不是Callable时才能工作,因为execute不能返回Future。
我认为在您的场景中,您应该构建未来对象,以便它也可以容纳异常内容。因此,如果出现异常,则构建错误消息对象。
答案 3 :(得分:1)
我的原始问题询问如何使用Java ExecutorService实现“健壮”线程异常处理。感谢Angelo和Greg指出异常处理如何与ExecutorService.submit()和Future.get()一起使用。我修改后的代码片段如下所示。我在这里学到的关键点是Future.get()捕获所有异常。如果线程被中断或取消,则会得到相应的异常,否则,异常将被包装并作为ExecutionException重新抛出。
import java.util.concurrent.Callable; import java.util.concurrent.CancellationException; import java.util.concurrent.ExecutionException; import java.util.concurrent.ExecutorService; import java.util.concurrent.Executors; import java.util.concurrent.Future; public class TestThreadFailure { public static void main(String[] args) { int size = 1; ExecutorService exec = Executors.newFixedThreadPool(size); ThreadFailTask worker = new ThreadFailTask(); Future result = exec.submit(worker); try { Integer value = result.get(); System.out.println("Result: " + value); } catch (ExecutionException ex) { System.out.println("Caught failure: " + ex.toString()); exec.shutdownNow(); return; } catch (InterruptedException iex) { System.out.println("Thread interrupted: " + iex.toString()); } catch (CancellationException cex) { System.out.println("Thread cancelled: " + cex.toString()); } exec.shutdownNow(); throw new RuntimeException("Did not catch failure !!"); } public static class ThreadFailTask implements Callable { @Override public Integer call() { int nbuf = 65536; double[][] buf = new double[nbuf][nbuf]; return new Integer((int) buf[0][0]); } } }
答案 4 :(得分:1)
我对其他答案没有太大的运气,因为我需要实际的异常实例本身,而不仅仅是打印的堆栈跟踪。对我而言,涉及问题“ThreadPoolExecutor#afterExecute()”的Why is UncaughtExceptionHandler not called by ExecutorService?的已接受答案有效。
请参阅以下示例代码:
List<Runnable> tasks = new LinkedList<>();
for (int i = 0; i < numThreads; ++i) {
Runnable task = new Runnable() {
@Override
public void run() {
throw new RuntimeException();
}
};
tasks.add(task);
}
Optional<Throwable> opEmpty = Optional.empty();
/*
* Use AtomicReference as a means of capturing the first thrown exception, since a
* spawned thread can't "throw" an exception to the parent thread.
*/
final AtomicReference<Optional<Throwable>> firstThrownException =
new AtomicReference<>(opEmpty);
/*
* Use new ThreadPoolExecutor instead of Executors.newFixedThreadPool() so
* that I can override afterExecute() for the purposes of throwing an
* exception from the test thread if a child thread fails.
*/
ExecutorService execSvc = new ThreadPoolExecutor(numThreads, numThreads,
0L, TimeUnit.MILLISECONDS, new LinkedBlockingQueue<Runnable>()) {
@Override
public void afterExecute(Runnable task, Throwable failureCause) {
if(failureCause == null) {
// The Runnable completed successfully.
return;
}
// only sets the first exception because it will only be empty on the first call.
firstThrownException.compareAndSet(Optional.<Throwable>empty(), Optional.of(failureCause));
}
};
for (Runnable task : tasks) {
execSvc.execute(task);
}
execSvc.shutdown();
execSvc.awaitTermination(1, TimeUnit.HOURS);
assertEquals(firstThrownException.get(), Optional.empty());
答案 5 :(得分:0)