Question

I have a class extending Thread class as follows.

public class RemoteDataSynchTaskProcessor extends Thread {
private GDRType taskGDR;

public RemoteDataSynchTaskProcessor(GDRType taskGDR) {
    this.taskGDR = taskGDR;
}

private void execute() {

    try {
        try {

            RemoteClient.runDataSynchTask(taskGDR);

        } catch (RemoteException re) {
            //.....
        } catch (NotBoundException e) {
            //......
        }
        processResult(false, "Success");
    } catch (DTSException de) {
    //....
    }

}

public void run() {
    Observer.incrementNumOfDataSyncyProcessor();

    try {
        execute();
    } finally {
        Observer.decrementNumOfDataSynchProcessor();
    }
}

After this thread’s execution is over there appears another thread named: Thread [pool-34-thread-1] (Running)

I have no idea why this thread appears and never terminates, and i end up with thousands of threads after a couple of day with pool number increasing every time ( e.g. Thread [pool-2500-thread-1])

If if debug and suspend execution of these generated threads, the stack is as follows:

Thread [pool-1-thread-1] (Suspended)    
Unsafe.park(boolean, long) line: not available [native method]  
LockSupport.park(Object) line: 156  
AbstractQueuedSynchronizer$ConditionObject.await() line: 1987   
LinkedBlockingQueue<E>.take() line: 399 
ThreadPoolExecutor.getTask() line: 947  
ThreadPoolExecutor$Worker.run() line: 907   
Thread.run() line: 662  

What are these threads, how can i get rid of them ?

EDIT: Seems like it has something to do with closed source thread pooling lib i am using (judging by the comments asking about executor service).

I use a library that provides ThreadPool capability. This is how i create a pool at the beginning of application startup.

    ThreadPool fetcherPool = new ThreadPool("TASK_FETCHER", ConfigParams.minimumTaskFetcherCount, ConfigParams.maximumTaskFetcherCount, new TaskFetcher());

ThreadPool.java seems to be extending Java’s ThreadGroup class, and ThreadFetcher is as follows, which actually triggers the thread.

public class TaskFetcher extends PooledMessageWorker //Pooled Message Worker extends ReacreatableThread

public void run() {
    Observer.incrementNumOfTaskFetcher();
    try {
        while (true) {
            try {
                if (Observer.isShutdowned()) {
                    break;
                }
                ControllerUtil.getInstance().waitIfPaused();

                ESDRType task = (ESDRType) TaskHandler.getTask();
                if (task != null) {
                    trigger(task);
                }
            } catch (Exception e) {
                LogUtil.error(logger, e.getMessage(), e);
            }
        }
    } finally {
        Observer.decrementNumOfTaskFetcher();
    }
}


private void trigger(ESDRType task) {
    try {
        GDRType[] messages = createMessages(esdr);
    for (GDRType message : messages) {
        RemoteDataSynchTaskProcessor remoteCDCMTaskProcessor = new      RemoteDataSynchTaskProcessor(message);
        remoteCDCMTaskProcessor.start(); //This thread causes a new thread appear after execution finishes
        LogUtil.debug(logger, "[distributeMessage()] [Message: ", message, "]... [OK]");    
    }
    } catch (DTSException exception) {
    // Logging stuff
    }
}

}

Answer 1

Sorry everyone, seems like it has nothing to do with that ThreadPool or RemoteDataSynchTaskProcessor.

I havent noticed another ExecutorService is created in a method that is called by ProcessResults() method in RemoteDataSynchTaskProcessor. And executor server is not shutdown after its job is down. My bad.

So if anyone has a problem like this one with unexplained thrad occurrences, check your code twice 😀 Look for the problem somewhere else 😛