Why does BoundedExecutor have to catch RejectedExecutionException? - java

As to BoundedExecutor in the book Java Concurrency in Practice, task submission has been throttled by semaphore. When would the underlying executor throw a RejectedExecutionException? Maybe when operating system runs out of threads?
public class BoundedExecutor {
private final Executor exec;
private final Semaphore semaphore;
public BoundedExecutor(Executor exec, int bound) {
this.exec = exec;
this.semaphore = new Semaphore(bound);
}
public void submitTask(final Runnable command) throws InterruptedException, RejectedExecutionException
{
semaphore.acquire();
try {
exec.execute(new Runnable() {
#Override public void run() {
try {
command.run();
} finally {
semaphore.release();
}
}
});
} catch (RejectedExecutionException e) {
semaphore.release();
throw e;
}
}
}

Part of the contract of Executor.execute() is that it can throw a RejectedExecutionException:
if this task cannot be accepted for execution
What that means for any given Executor implementation is up to the discretion of that implementation. I could create an OnlyOnSundaysExecutor that rejects tasks unless the current day of the week is Sunday. You would have to check the documentation for the various Executor implementations to see under what circumstances they would throw a RejectedExecutionException exception.

Regardless of the circumstances under which an exception might be thrown, it is important to keep your application in a consistent state, even when exceptions occur.
Here, the semaphore that has been acquired should always get released. For most resources that should get released under all circumstances, you would use a try {} finally { /* release action */ } construct to ensure the release even in the exceptional case, but here, we have the special situation that the release action should be performed only in the exceptional case, as in the successful case, the submitted Runnable will perform the release action (note that within the Runnable, finally is indeed being used).
So the code is supposed to handle the case when a RejectedExecutionException is thrown. We might want to do that for every RuntimeException or Error, but the problem is, RejectedExecutionException is the only exception type where we know for sure that the runnable will never get executed. For all other kinds of exceptions, it is possible that it will still run.
To make the cleanup safe, you would need another atomic toggle:
public void submitTask(final Runnable command) throws InterruptedException {
AtomicBoolean proceed = new AtomicBoolean(true);
semaphore.acquire();
try {
exec.execute(new Runnable() {
#Override public void run() {
if(proceed.compareAndSet(true, false)) try {
command.run();
} finally {
semaphore.release();
}
}
});
} catch(Throwable e) {
if(proceed.compareAndSet(true, false)) semaphore.release();
throw e;
}
}
Now, in every exceptional case, an attempt to release the semaphore will be made, unless the Runnable flagged that it is already running. Or the Runnable will not proceed if it detects that the semaphore has been released due to an exception in the submitting code.
This is, of course, more complicated than the book example and possibly distracting from the original intent of the example. Also, it uses the Java 7 feature of being able to easily catch and re-throw all possible exceptions of the try block. This wasn’t available when the book was written.

Related

How to reset interrupted flag in executorservice.invokeAll()

I've learnt that it is the good practice to set Thread.interrupter() flag back to true after catching a InterruptedException.
However when I call executorService.invokeAll() and put a try clause around it, by catching the InterruptedException, will Thread.currentThread().interrupt() correctly flag the interrupted thread, or the main thread of my controller/runner?
public class Controller {
public void run() {
final List<Task> tasks = request.collectTasks()
try {
executorService.invokeAll(tasks);
} catch (InterruptedException e) {
message = "interrupted";
Thread.currentThread().interrupt();
throw new InternalFailureException(message);
}
}
}
class Task implements Callable<String> {
#Override
public String call() {
return taskId;
}
}
Overall is this the correct way to reset the interrupted flag?
In this code example you have given
try {
executorService.invokeAll(tasks);
} catch (InterruptedException e) {
message = "interrupted";
Thread.currentThread().interrupt();
throw new InternalFailureException(message);
}
whoever invokes Thread.currentThread().interrupt(); will set the flag of that thread in this example it will be the thread executing your Controller#run() method.
Ideally with executors it is the responsibility of Task or Runnable which gets scheduled over the executor to handle this flag properly.
Since it is a checked exception already - it will not give you an opportunity to decorate the task that you are about to submit inside executor so you have no easy control over it.

Stopping a Thread / Threads calls interrupt on itself after crash?

I am currently running a Thread from a Service to do some background work.
Now there is the possibility that the Thread crashes or I want to
interrupt the thread from the Service. So how am I supposed to:
stop the Thread realiable, (hard)
catch exceptions and call the Service about the crash
handle InterruptedException if interrupted while sleep()
is Thread.isInterrupted a good way to detect if the Thread stopped?
What I have done so far is the following:
#Override
public void run() {
try {
while (!Thread.currentThread().isInterrupted()) {
doMyBackgroundWork();
sleep();
}
}catch(Exception e){
ExceptionHandler.logAndSendException(e);
Thread.currentThread().interrupt();
if(crashedListener != null){
crashedListener.onThreadCrashed();
}
}
LOG.i("Thread stops now.");
}
private void sleep() {
try {
sleep(frequency);
} catch (InterruptedException e) {
//what to do here? it can happen because I stopped it myself
}
}
So at first I am running my Thread until it gets interrupted.
If any exception occurs, I want to start a new Thread, therefore
my Service implements a listener interface and I call it, once an
Exception is thrown. I know that catching everything is discouraged,
but I need to know if the Thread stops, without polling Thread.isAlive()
all the time.
Additionally to my four questions above:
is my code reliable and does what I need?
is it ok to call interrupt on the Thread itself?
Thanks!
You are not actually interrupting your own thread because the catch block is outside of the while loop. Therefore, any exception would stop execution immediately.
Interruption is essentially just a request (usually from another thread) to stop doing what you are doing. The thread is free to ignore it and keep doing what it is doing. Normally you have to throw an exception in response to an interrupt, or stop execution some other way such as just breaking from the loop (you need this around the //what to do here? comment). It so happens that some library methods are "responsive to interruption" meaning they will throw an exception if the thread is ever interrupted, such as Thread.sleep(), which you will most likely have in your sleep call.
I recommend picking Java Concurrency In Practice. Among the excellent concurrency material, there is a chapter on interrupts which is very helpful.
EDIT:
I would remove the code where you interrupt your own thread. You will also need to rethrow the InterruptedException as a runtime exception to get out of the execution loop. Usually people will create a new Exception that extends RuntimeException that is something like MyInterruptedException. You can then add it to the catch block around your loop so that you know when the thread was interrupted vs execution failed.
As a general example you can do something like this:
public void run() {
try {
while (true) {
// check for interrupts in the loop, or somewhere in the work method
if (Thread.interrupted()) {
throw new MyInterruptedException("Important thread interrupted.");
}
doMyBackgroundWork();
sleep();
}
}
catch(Exception e){
ExceptionHandler.logAndSendException(e);
if(crashedListener != null){
crashedListener.onThreadCrashed();
}
}
catch(MyInterruptedException i) {
LOG.i("Execution stopping because of interrupt.");
}
}
private void sleep() {
try {
sleep(frequency);
} catch (InterruptedException e) {
throw new MyInterrptedException(e);
}
}
we have a nice and effective method called stop()(Thread.stop(void):void) which is deprecated, but it works and it's lovely.
Note that stop() throws ThreadDeath at the target thread which is not an exception(and it could any other throwable too), but an Error, so your code will not catch any signal about this.
public void run() {
try {
while (<<using_a_volatile_bool_type_is_better>>) {
...
}
}catch(Throwable t){/**/}/*use throwable instead of exception.*/}
}
Beside dear friend stop() we also have pause() method too, and it really pauses the target thread.
Not just one solution out there, but if it's really critical to keep thread run and run the emergency(or itself) just after any crash, you may run it as a separately app/process, plus get progress status(if any) that ensures you the target thread/app is not freezed(blocked,...)

Resume interrupted thread

I want to resume the work of interrupted thread,please let me know some possible solutions for the same.
class RunnableDemo implements Runnable
{
public void run()
{
while(thread.isInterrupted())
{
try{}
catch(Exception e){ //exception caught}
}
}
}
If exception is caught, thread is interrupted, but even though exception is caught, i want thread to continue its work, so please suggest me some way to overcome this issue.
Thanks in advance
Thread interruption is something you choose to obey when writing a thread. So if you don't want your thread to be interrupted, don't check the interrupted status and continue regardless.
The only time you'll need try/catch statements (with respect to thread interruption) is when calling blocking methods that throw InterruptedException. Then you'll need to avoid letting that exception stop your thread's work.
Of course... you should give some thought about whether this is a suitable way to behave. Thread interruption is a helpful thing and choosing not to adhere to it can be annoying to users of your code.
I have written a reusable code for getting this feature where thread can be pause and resume. Please find the code below. Your can extend PausableTask and override task() method:
public abstract class PausableTask implements Runnable{
private ExecutorService executor = Executors.newSingleThreadExecutor();
private Future<?> publisher;
protected volatile int counter;
private void someJob() {
System.out.println("Job Done :- " + counter);
}
abstract void task();
#Override
public void run() {
while(!Thread.currentThread().interrupted()){
task();
}
}
public void start(){
publisher = executor.submit(this);
}
public void pause() {
publisher.cancel(true);
}
public void resume() {
start();
}
public void stop() {
executor.shutdownNow();
}
}
Hope this helps. For further details check this link or give me shout in comment section.
http://handling-thread.blogspot.co.uk/2012/05/pause-and-resume-thread.html
A thread get's interrupted only if someone called the interrupt() method of that thread and not because some other random exception was thrown while running your thread as you are thinking.
When the thread's interrupted() method is called, InterruptedException will be thrown in the thread if the thread is in the middle of a blocking operation (eg. IO read).
When the InterruptedException is thrown you should know that the interrupted status is cleared, so the next time you call isInterrupted() in your thread will give you false (even though you just cauth the InterruptedException)
Have this in mind while coding your threads. And if you don't understand what I am talking about stop coding multithreading and go read some books about concurrency in java.
One caveat: If your thread handles an InterruptedException while in a call to a third-party library, then you won't necessarily know how the library reacted to it (i.e., did it leave the library objects in a state when it makes sense for your program to continue using them?)
Some developers (including some library developers) mistakenly assume that an interrupt means, "shut down the program," and all they worry about is closing files, etc.; and not so much about whether the library can continue to be used.
Try it and see, but if you're writing code to control a spacecraft or a nuclear reactor or something, then you may want to do a little extra work to really find out what the library does.
As others already stated, usually interruption is the proper way to cancel a task. If you really need to implement a non-cancellable task, at least make sure to restore the interrupted-state of the thread when you're done with your non-interruptible work:
public void run() {
boolean interrupted = false;
try {
while (true) {
try {
callInterruptibleMethod();
} catch (InterruptedException ex) {
interrupted = true;
// fall through and retry
}
}
} finally {
if (interrupted) {
// restore interruption state
Thread.currentThread().interrupt();
}
}
}
(From book: Java Concurrency in Practice)

Should I Thread.currentThread.interrupt() before I throw an exception back?

I am implementing an interface which throws IOException. In my implementation, I call another method which can block, and therefore throw InterruptedException.
Context:
I want to end the treatment if I am interrupted;
this is not a thread I created myself.
My current idea is to do as such (skeleton code):
#Override
public void implementedMethod()
throws IOException
{
try {
methodThatBlocks();
} catch (InterruptedException ignored) {
Thread.currentThread().interrupt();
throw new IOException();
}
}
is that the correct way? Or should I just throw and not .interrupt()?
Yes, you should call interrupt() to let the calling code know that the thread has been interrupted. If you don't do it, since the InterruptedException clears it, the calling code will have no way to know about the interruption and won't stop running although it should.
Let me quote Java Concurrency in Practice:
Restore the interrupt. Sometimes you cannot throw InterruptedException, for instance when your code is part of a Runnable. In these situations, you must catch InterruptedException and restore the interrupted status by calling interrupt on the current thread, so that code higher up the call stack can see that an interrupt was issued,
as demonstrated in Listing 5.10.
public class TaskRunnable implements Runnable {
BlockingQueue<Task> queue;
...
public void run() {
try {
processTask(queue.take());
} catch (InterruptedException e) {
// restore interrupted status
Thread.currentThread().interrupt();
}
}
}
Of course not. The thread is already unblocked and running. In the code that is about to throw an exception. What would be the point?

Thread Blocks in server mode

This method notifes an event loop to start processing a message. However, if the event loop is already processing a message then, this method blocks until it receives a notification of completed event processing (which is triggered at the end of the event loop).
public void processEvent(EventMessage request) throws Exception {
System.out.println("processEvent");
if (processingEvent) {
synchronized (eventCompleted) {
System.out.println("processEvent: Wait for Event to completed");
eventCompleted.wait();
System.out.println("processEvent: Event completed");
}
}
myRequest = request;
processingEvent = true;
synchronized (eventReady) {
eventReady.notifyAll();
}
}
This works in client mode. If I switch to server mode and the time spent in the event loop processing the message is too quick, then the method above blocks forever waiting for the event to completed. For some reason the event complete notification is sent after the processingEvent check and before the eventCompleted.wait(). It makes no difference if I remove the output statements. I can not repeat the same problem in client mode.
Why does this only happen in server mode and what can I do to prevent this happening?
Here is the eventReady wait and eventCompleted notification:
public void run() {
try {
while (true) {
try {
synchronized (eventReady) {
eventReady.wait();
}
nx.processEvent(myRequest, myResultSet);
if (processingEvent > 0) {
notifyInterface.notifyEventComplete(myRequest);
}
} catch (InterruptedException e) {
throw e;
} catch (Exception e) {
notifyInterface.notifyException(e, myRequest);
} finally {
processingEvent--;
synchronized (eventCompleted) {
eventCompleted.notifyAll();
}
}
} // End of while loop
} catch (InterruptedException Ignore) {
} finally {
me = null;
}
Here is revised code which seems to work without the deadlock problem - which BTW happened in client mode randomely after about 300 events.
private BlockingQueue<EventMessage> queue = new SynchronousQueue<EventMessage>();
public void processEvent(EventMessage request) throws Exception {
System.out.println("processEvent");
queue.put(request);
}
public void run() {
try {
while (true) {
EventMessage request = null;
try {
request = queue.take();
processingEvent = true;
nx.processEvent(request, myResultSet);
notifyInterface.notifyEventComplete(request);
} catch (InterruptedException e) {
throw e;
} catch (Exception e) {
notifyInterface.notifyException(e, request);
} finally {
if (processingEvent) {
synchronized (eventCompleted) {
processingEvent = false;
eventCompleted.notifyAll();
}
}
}
} // End of while loop
} catch (InterruptedException Ignore) {
} finally {
me = null;
}
}
If you call notifyAll and no thread is wait()ing, the notify is lost.
The correct approach is to always change a state, inside the synchronized block, when calling notify() and always check that state, inside the synchronized block, before calling wait().
Also your use of processingEvent doesn't appear to be thread safe.
Can you provide the code which waits on eventReady and notifies eventCompleted?
Your program can happen to work if your speed up or slow down your application just right e.g. if you use -client, but if you use a different machine, JVM or JVM options it can fail.
There are a number of race conditions in your code. Even declaring processingEvent volatile or using an AtomicBoolean won't help. I would recommend using a SynchronousQueue which will block the event until the processer is ready for it. Something like:
private final BlockingQueue<Request> queue = new SynchronousQueue<Request>();
...
// this will block until the processor dequeues it
queue.put(request);
Then the event processor does:
while (!done) {
// this will block until an event is put-ed to the queue
Request request = queue.take();
process the event ...
}
Only one request will be processed at once and all of the synchronization, etc. will be handled by the SynchronousQueue.
If processingEvent isn't declared volatile or accessed from within a synchronized block then updates made by one thread may not become visible to other threads immediately. It's not clear from your code whether this is the case, though.
The "server" VM is optimised for speed (at the expense of startup time and memory usage) which could be the reason why you didn't encounter this problem when using the "client" VM.
There is a race condition in your code that may be exasperated by using the server VM, and if processingEvent is not volatile then perhaps certain optimizations made by the server VM or its environment are further influencing the problem.
The problem with your code (assuming this method is accessed by multiple threads concurrently) is that between your check of processingEvent and eventCompleted.wait(), another thread can already notify and (I assume) set processingEvent to false.
The simplest solution to your blocking problem is to not try to manage it yourself, and just let the JVM do it by using a shared lock (if you only want to process one event at a time). So you could just synchronize the entire method, for instance, and not worry about it.
A second simple solution is to use a SynchronousQueue (this is the type of situation it is designed for) for your event passing; or if you have more executing threads and want more than 1 element in the queue at a time then you can use an ArrayBlockingQueue instead. Eg:
private SynchronousQueue<EventMessage> queue = new SynchronousQueue<EventMessage>();
public void addEvent(EventMessage request) throws Exception
{
System.out.println("Adding event");
queue.put(request);
}
public void processNextEvent()
{
EventMessage request = queue.take();
processMyEvent(request);
}
// Your queue executing thread
public void run()
{
while(!terminated)
{
processNextEvent();
}
}

Categories