I am performing some operations that are time sensitive and have a timeout associated with them.
This timeout mechanism is implemented using the Java Callable class.
The problem is that within the callable instance I execute an asynchronous task, with an anonymous interface implementation (listener).
My problem is that when the timeout triggers and the callable is cancelled. The async callback still executes and corrupts the state of my program.
How do I prevent these callbacks from firing? Do I just include a boolean specififying whether or not the timeout has ocurred or is there another way of achieving this please?
Thanks.
Code reference:
Callable<Object> callableTransaction = new Callable<Object>() {
#Override
public Object call() throws Exception {
Callback callback = new Callback() {
// Do stuff here and change program state.
};
performAsyncOperation(callback);
return ActionProcessor.OPERATION_COMPLETE;
}
};
Why don't you remove performAsyncOperation and perform the operation synchronously inside the call method.
And then, if you need any async operation, you can invoke your callableTransaction using performAsyncOperation or executor services.
Wrap your Callable in a java.util.concurrent.FutureTask... call task.cancel(true) when you want to cancel things. If your Callable is in a blocking operation (I/O, for example), then an exception will be thrown marking the interruption. Otherwise, your thread will need to check using Thread.isInterrupted() periodically to see if it should continue or abort.
Yes, this is mostly the same as including your own boolean flag, but you do get the benefit of blocking operations getting interrupted as well.
You have the right idea: Use a boolean flag to track whether the Callable has been interrupted. (I'm assuming that when you cancel the Callable, you are specifying that it should be interrupted.)
Use a concurrent class like CountDownLatch to make your Callable behave synchronously:
public Object call() throws Exception {
final AtomicBoolean aborted = new AtomicBoolean();
final CountDownLatch latch = new CountDownLatch(1);
Callback callback = new Callback() {
if (aborted.get()) {
return;
}
// Do stuff here and change program state.
latch.countDown(); // Tell Callable we're done.
};
performAsyncOperation(callback);
try {
latch.await(); // Wait for callback to finish.
} catch (InterruptedException e) {
aborted.set(true);
}
return ActionProcessor.OPERATION_COMPLETE;
}
I am learning to use ExectorService to pool threads and send out tasks. I have a simple program below
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.TimeUnit;
class Processor implements Runnable {
private int id;
public Processor(int id) {
this.id = id;
}
public void run() {
System.out.println("Starting: " + id);
try {
Thread.sleep(5000);
} catch (InterruptedException e) {
System.out.println("sorry, being interupted, good bye!");
System.out.println("Interrupted " + Thread.currentThread().getName());
e.printStackTrace();
}
System.out.println("Completed: " + id);
}
}
public class ExecutorExample {
public static void main(String[] args) {
Boolean isCompleted = false;
ExecutorService executor = Executors.newFixedThreadPool(2);
for (int i = 0; i < 5; i++) {
executor.execute(new Processor(i));
}
//executor does not accept any more tasks but the submitted tasks continue
executor.shutdown();
System.out.println("All tasks submitted.");
try {
//wait for the exectutor to terminate normally, which will return true
//if timeout happens, returns false, but this does NOT interrupt the threads
isCompleted = executor.awaitTermination(100, TimeUnit.SECONDS);
//this will interrupt thread it manages. catch the interrupted exception in the threads
//If not, threads will run forever and executor will never be able to shutdown.
executor.shutdownNow();
} catch (InterruptedException e) {
}
if (isCompleted) {
System.out.println("All tasks completed.");
} else {
System.out.println("Timeout " + Thread.currentThread().getName());
}
}
}
It does nothing fancy, but creates two threads and submits 5 tasks in total. After each thread completes its task, it takes the next one,
In the code above, I use executor.submit. I also changed to executor.execute. But I do not see any difference in the output. In what way are the submit and execute methods different?
This what the API says
Method submit extends base method Executor.execute(java.lang.Runnable) by creating and returning a Future that can be used to cancel execution and/or wait for completion. Methods invokeAny and invokeAll perform the most commonly useful forms of bulk execution, executing a collection of tasks and then waiting for at least one, or all, to complete. (Class ExecutorCompletionService can be used to write customized variants of these methods.)
But it's not clear to me as what it exactly means?
As you see from the JavaDoc execute(Runnable) does not return anything.
However, submit(Callable<T>) returns a Future object which allows a way for you to programatically cancel the running thread later as well as get the T that is returned when the Callable completes. See JavaDoc of Future for more details
Future<?> future = executor.submit(longRunningJob);
...
//long running job is taking too long
future.cancel(true);
Moreover,
if future.get() == null and doesn't throw any exception then Runnable executed successfully
The difference is that execute simply starts the task without any further ado, whereas submit returns a Future object to manage the task. You can do the following things with the Future object:
Cancel the task prematurely, with the cancel method.
Wait for the task to finish executing, with get.
The Future interface is more useful if you submit a Callable to the pool. The return value of the call method will be returned when you call Future.get. If you don't maintain a reference to the Future, there is no difference.
execute: Use it for fire and forget calls
submit: Use it to inspect the result of method call and take appropriate action on Future objected returned by the call
Major difference: Exception handling
submit() hides un-handled Exception in framework itself.
execute() throws un-handled Exception.
Solution for handling Exceptions with submit()
Wrap your Callable or Runnable code in try{} catch{} block
OR
Keep future.get() call in try{} catch{} block
OR
implement your own ThreadPoolExecutor and override afterExecute method
Regarding tour other queries on
invokeAll:
Executes the given tasks, returning a list of Futures holding their status and results when all complete or the timeout expires, whichever happens first.
invokeAny:
Executes the given tasks, returning the result of one that has completed successfully (i.e., without throwing an exception), if any do before the given timeout elapses.
Use invokeAll if you want to wait for all submitted tasks to complete.
Use invokeAny if you are looking for successful completion of one task out of N submitted tasks. In this case, tasks in progress will be cancelled if one of the tasks completes successfully.
Related post with code example:
Choose between ExecutorService's submit and ExecutorService's execute
A main difference between the submit() and execute() method is that ExecuterService.submit()can return result of computation because it has a return type of Future, but execute() method cannot return anything because it's return type is void. The core interface in Java 1.5's Executor framework is the Executor interface which defines the execute(Runnable task) method, whose primary purpose is to separate the task from its execution.
Any task submitted to Executor can be executed by the same thread, a worker thread from a thread pool or any other thread.
On the other hand, submit() method is defined in the ExecutorService interface which is a sub-interface of Executor and adds the functionality of terminating the thread pool, along with adding submit() method which can accept a Callable task and return a result of computation.
Similarities between the execute() and submit() as well:
Both submit() and execute() methods are used to submit a task to Executor framework for asynchronous execution.
Both submit() and execute() can accept a Runnable task.
You can access submit() and execute() from the ExecutorService interface because it also extends the Executor interface which declares the execute() method.
Apart from the fact that submit() method can return output and execute() cannot, following are other notable differences between these two key methods of Executor framework of Java 5.
The submit() can accept both Runnable and Callable task but execute() can only accept the Runnable task.
The submit() method is declared in ExecutorService interface while execute() method is declared in the Executor interface.
The return type of submit() method is a Future object but return type of execute() method is void.
If you check the source code, you will see that submit is sort of a wrapper on execute
public Future<?> submit(Runnable task) {
if (task == null) throw new NullPointerException();
RunnableFuture<Void> ftask = newTaskFor(task, null);
execute(ftask);
return ftask;
}
Submit - Returns Future object, which can be used to check result of submitted task. Can be used to cancel or to check isDone etc.
Execute - doesn't return anything.
The execute(Runnable command) is the implemented method from Interface Executor. It means just execute the command and gets nothing returned.
ExecutorService has its own methods for starting tasks: submit, invokeAny and invokeAll all of which have Callable instances as their main targets. Though there're methods having Runnable as input, actulaly Runnable will be adapted to Callable in the method. why Callable? Because we can get a Future<T> result after the task is submitted.
But when you transform a Runnable to a Callable, result you get is just the value you pass:
static final class RunnableAdapter<T> implements Callable<T> {
final Runnable task;
final T result;
RunnableAdapter(Runnable task, T result) {
this.task = task;
this.result = result;
}
public T call() {
task.run();
return result;
}
}
So, what's the point that we pass a Runnable to submit instead of just getting the result when the task is finished? Because there's a method which has only Runnable as parameter without a particular result.
Read the javadoc of Future:
If you would like to use a Future for the sake of cancellability but not provide a usable result, you can declare types of the form Future<?> and return null as a result of the underlying task.
So, if you just want to execute a Runnable task without any value returned, you can use execute().
if you want to run a Callable task, or
if you want to run a Runnable task with a specified result as the completion symbol, or
if you want to run a task and have the ability to cancel it,
you should use submit().
On top of previous responses, i.e.
execute(..) runs the task and forget about it
submit(...) returns a future;
The main advantage with the future is that you can establish a timeout. This might come very handy if you have an executor with a limited number of threads and your executions are taking forever, it will not hang the process.
Example 1: hangs forever and fills the executor
ExecutorService executor = Executors.newFixedThreadPool(2);
for (int i=0; i < 5; i++) {
executor.execute(() -> {
while (true) {
System.out.println("Running...")
Thread.sleep(Long.MAX_VALUE)
}
});
}
Your output will be (i.e. only 2 and it gets stuck):
Running...
Running...
On the other hand, you can use submit and add a timeout:
ExecutorService executor = Executors.newFixedThreadPool(2);
for (int i=0; i < 5; i++) {
Future future = executor.submit(() -> {
while (true) {
System.out.println("Running...");
Thread.sleep(Long.MAX_VALUE);
}
});
try {
future.get(1, TimeUnit.SECONDS);
} catch (Exception e) {
if (!future.isDone()) {
System.out.println("Oops: " + e.getClass().getSimpleName());
future.cancel(true);
}
}
}
The output will look like this (notice that the executor does not get stuck, but you need to manually cancel the future):
Running...
Oops: TimeoutException
Running...
Oops: TimeoutException
Running...
Oops: TimeoutException
Running...
Oops: TimeoutException
Running...
Oops: TimeoutException
basically both calls execute,if u want future object you shall call submit() method
here from the doc
public <T> Future<T> submit(Callable<T> task) {
if (task == null) throw new NullPointerException();
RunnableFuture<T> ftask = newTaskFor(task);
execute(ftask);
return ftask;
}
public <T> Future<T> submit(Runnable task, T result) {
if (task == null) throw new NullPointerException();
RunnableFuture<T> ftask = newTaskFor(task, result);
execute(ftask);
return ftask;
}
as you can see java really has no way to start a thread other than calling run() method, IMO. since i also found that Callable.call() method is called inside run() method. hence if the object is callable it would still call run() method, which inturn would call call() method
from doc.
public void run() {
if (state != NEW ||
!UNSAFE.compareAndSwapObject(this, runnerOffset,
null, Thread.currentThread()))
return;
try {
Callable<V> c = callable;
if (c != null && state == NEW) {
V result;
boolean ran;
try {
result = c.call();
ran = true;
} catch (Throwable ex) {
result = null;
ran = false;
setException(ex);
}
if (ran)
set(result);
}
} finally {
// runner must be non-null until state is settled to
// prevent concurrent calls to run()
runner = null;
// state must be re-read after nulling runner to prevent
// leaked interrupts
int s = state;
if (s >= INTERRUPTING)
handlePossibleCancellationInterrupt(s);
}
}
How to go about implementing a "fire and forget" behavior with java.util.concurrency? I tried:
ExecutorService executor = Executors.newSingleThreadExecutor();
public void push(Callable<Boolean> task) {
Future<Boolean> future = executor.submit(task);
future.get(timeout, timeoutUnit);
}
but the get() is blocking until completion. The push() caller is not interested in the result of the task.
Don't call get(). submit() is enough to start the Callable in a Thread. You can pass around the Future and call its get() when you are ready to get the result (if ever).
The Javadoc states
Submits a value-returning task for execution and returns a Future
representing the pending results of the task. The Future's get method
will return the task's result upon successful completion.
If you would like to immediately block waiting for a task, you can use
constructions of the form result = exec.submit(aCallable).get();
So just don't call get().
dont need the future object if you have to forget it i.e. dont care about the future of the thread. :)
ExecutorService executor = Executors.newSingleThreadExecutor();
public void push(Callable<Boolean> task) {
executor.submit(task);
}
or if you need to use future some time later than :-
ExecutorService executor = Executors.newSingleThreadExecutor();
public void push(Callable<Boolean> task) {
someCollection.add(executor.submit(task)); // use futures later
}
or just use execute from Executor and pass Runnable, if you dont intent to get the future at all.
ExecutorService executor = Executors.newSingleThreadExecutor();
public void push(Runnable task) {
executor.execute(task);
}
// execute will invoke default exceptional handler in case of exception, that can be lost in case if you dont get hold of futures in submit method.
Hey there i currently have a problem with my android app. I´m starting an extra thread via
implementing the Excecutor Interface:
class Flasher implements Executor {
Thread t;
public void execute(Runnable r) {
t = new Thread(r){
};
t.start();
}
}
I start my Runnable like this:
flasherThread.execute(flashRunnable);
but how can i stop it?
Ok, this is just the very basic threading 101, but let there be another example:
Old-school threading:
class MyTask implements Runnable {
public volatile boolean doTerminate;
public void run() {
while ( ! doTerminate ) {
// do some work, like:
on();
Thread.sleep(1000);
off();
Thread.sleep(1000);
}
}
}
then
MyTask task = new MyTask();
Thread thread = new Thread( task );
thread.start();
// let task run for a while...
task.doTerminate = true;
// wait for task/thread to terminate:
thread.join();
// task and thread finished executing
Edit:
Just stumbled upon this very informative Article about how to stop threads.
Not sure that implementing Executor is a good idea. I would rather use one of the executors Java provides. They allow you to control your Runnable instance via Future interface:
ExecutorService executorService = Executors.newSingleThreadExecutor();
Future<?> future = executorService.submit(flashRunnable);
...
future.cancel(true);
Also make sure you free resources that ExecutorService is consuming by calling executorService.shutdown() when your program does not need asynchronous execution anymore.
Instead of implementing your own Executor, you should look at ExecutorService. ExecutorService has a shutdown method which:
Initiates an orderly shutdown in which previously submitted tasks are executed, but no new tasks will be accepted.
I would suggest to use the ExecutorService, along with the Furure object, which gives you control over the thread that is being created by the executor. Like the following example
ExecutorService executor = Executors.newSingleThreadExecutor();
Future future = executor.submit(runnabale);
try {
future.get(2, TimeUnit.SECONDS);
} catch (TimeoutException ex) {
Log.warn("Time out expired");
} finally {
if(!future.isDone()&&(!future.isCancelled()))
future.cancel(true);
executor.shutdownNow();
}
This code says that the runnable will be forced to terminate after 2 seconds. Of course, you can handle your Future ojbect as you wish and terminate it according to your requierements
I am following up an interesting question on so, on usage of ScheduledThreadPoolExecutor for some repeating task.
Scheduling this object returns a ScheduledFuture object which one can use to cancel the next run of the task.
One thing to note here is the task itself is completely decoupled from the schedule--
ScheduledExecutorService executor = new ScheduledThreadPoolExecutor(1);
ScheduledFuture nextSchedule =
executor.schedule(task, 60000, TimeUnit.MILLISECONDS);
where-
SomeTask task = new SomeTask();
So the task itself is not aware of the schedule. Please enlighten if there is a way to get the task to cancel and create a new schedule for itself.
Thanks
There's no reason why the task cannot reference the ScheduledExecutorService and schedule itself to run again if required:
// (Need to make variable final *if* it is a local (method) variable.)
final ScheduledExecutorService execService = Executors.newSingleThreadScheduledExecutor();
// Create re-usable Callable. In cases where the Callable has state
// we may need to create a new instance each time depending on requirements.
Callable<Void> task = new Callable() {
public Void call() {
try {
doSomeProcessing();
} finally {
// Schedule same task to run again (even if processing fails).
execService.schedule(this, 1, TimeUnit.SECONDS);
}
}
}
Pass the executor to the task, so that it can make manipulations with it:
SomeTask task = new SomeTask(executor);