Thread safety issue while timing out the thread for each bundle - java

I am working on a project in which I will have different Bundles/Models. Let's take an example, Suppose I have 4 bundles and each of those bundles will have a method name process.
Below are the things, I am supposed to do-
I need to call all those 4 Bundles process method in parallel using multithread and process method in each bundle will return me a map and then write this map into the database in that same thread or whatever is the best approach to do (I am not sure on this which is the right way to go).
And also I want to have some sort of timeout feature enabled at the thread level. Meaning if any Bundle is taking lot of time to execute, then that Bundle thread should get timeout and log as an error stating that this particular bundle got timeout bcoz it was taking lot of time.
The following attempt that I have done is most probably flawed and error handling is by no means complete. Can anybody guide me what I am supposed to do in the error handling cases as well?
Below is my method which will call process method of all the bundles in a multithreaded way.
public void processEvents(final Map<String, Object> eventData) {
ExecutorService pool = Executors.newFixedThreadPool(5);
List<ProcessBundleHolderEntry> entries = new ArrayList<ProcessBundleHolderEntry>();
Map<String, String> outputs = (Map<String, String>)eventData.get(BConstants.EVENT_HOLDER);
for (BundleRegistration.BundlesHolderEntry entry : BundleRegistration.getInstance()) {
ProcessBundleHolderEntry processBundleHolderEntry = new ProcessBundleHolderEntry(entry, outputs);
entries.add(processBundleHolderEntry);
}
try {
List<Future<Object>> futures = pool.invokeAll(entries, 30, TimeUnit.SECONDS);
for (int i = 0; i < futures.size(); i++) {
// This works since the list of future objects are in the
// same sequential order as the list of entries
Future<Object> future = futures.get(i);
ProcessBundleHolderEntry entry = entries.get(i);
if (!future.isDone()) {
// log error for this entry
}
}
} catch (InterruptedException e) {
// handle this exception!
}
}
Secondly, an implementation of Callable for your threads:
public class ProcessBundleHolderEntry implements Callable {
private BundleRegistration.BundlesHolderEntry entry;
private Map<String, String> outputs;
public ProcessBundleHolderEntry(BundleRegistration.BundlesHolderEntry entry, Map<String, String> outputs) {
this.entry = entry;
this.outputs = outputs;
}
public Object call() throws Exception {
final Map<String, String> response = entry.getPlugin().process(outputs);
// write to the database.
System.out.println(response);
return response;
}
}
Can anyone tell me whether there is any problem with the above approach or is there any better and efficient way of doing the same thing? I am not sure whether there is any thread safety issue as well.
Any help will be appreciated on this.

The only shared object in your code is eventData: as long as it is not modified while this method is running (or if the map and its content is thread safe and changes are safely published) you should be fine.
Regarding exception handling of your tasks, you typically do:
try {
future.get();
} catch (ExecutionException e) {
Throwable exceptionInFuture = e.getCause();
//throw, log or whatever is appropriate
}
Regarding the interrupted exception: it means the thread in which you are executing the method has been interrupted. What you need to do depends on your use case, but you should generally stop what you are doing, so something like:
} catch (InterruptedException e) {
pool.shutdownNow(); //cancels the tasks
//restore interrupted flag and exit
Thread.currentThread.interrupt();
//or rethrow the exception
throw e;
}
Note: the purpose of thread pools is to be reused - you should declare the executor service as an (private final) instance variable rather than creating one every time the processEvents method is called.

Related

Java FutureTask - Multithreaded call to get()

I have the following two methods in a class:
private MyDef myDef;
private FutureTask<MyDef> defFutureTask;
public synchronized void periodEviction() {
myDef = null;
}
public MyDef loadMyItems() {
// if it's not ready use a future - it will block until the results are ready
if (this.myDef == null) { // this will still not be thread safe
Callable<MyDef> callableDef = ()->{ return this.loadFromDatabase(); };
FutureTask<MyDef> defTask = new FutureTask<>(callableDef);
this.defFutureTask = defTask;
defFutureTask.run();
}
try {
// wait until's it's ready
this.myDef = this.qDefFuture.get();
} catch(InterruptedException e) {
log.error(this.getClass(), "Interrupted whilst getting future..");
} catch(ExecutionException e) {
log.error(this.getClass(), "Error when executing callable future");
}
return this.myDef;
}
I wanted to do the following:
1) Do a cache eviction using periodEviction() every one hour or so.
2) Otherwise, use the cached value when db loading is done.
I believe I have misunderstood Java future as I couldn't answer the question, "What happens when Thread A,B,and C all are calling loadMyItems() at the same time?"
So does this mean without something like an executor, this implementation is still not thread safe?
An even simpler approach is to not cache the object at all but just retain the Future.
private CompletableFuture<MyDef> defFuture;
public synchronized void periodEviction() {
// evict by triggering the request anew
defFuture = CompletableFuture.supplyAsync(this::loadFromDatabase);
}
public synchronized Optional<MyDef> loadMyItems() {
try {
return Optional.of(this.defFuture.get());
} catch(InterruptedException e) {
log.error(this.getClass(), "Interrupted whilst getting future..");
} catch(ExecutionException e) {
log.error(this.getClass(), "Error when executing callable future");
}
return Optional.empty();
}
With the caveat that this will trigger the database query every eviction period rather than on demand.
A super simple approach would be to declare loadMyItems as synchronized. But if the class has other methods that access myDef, you would have to declare those synchronized too. Sometimes this results in very coarse-grained locking and slower performance.
If you're looking for the cleanest/fastest code, instead of declaring periodEviction as synchronized, declare myDef as an AtomicReference:
private final AtomicReference<MyDef> myDef = new AtomicReference<>();
Then the body of periodEviction is:
synchronized (myDef) {
myDef.set(null);
}
And the body of loadMyItems is:
synchronized (myDef) {
if (myDef.get() == null) {
// perform initialization steps, ending with:
myDef.set(this.qDefFuture.get());
}
return myDef.get();
}
If many threads call loadMyItems at the same time, myDef will only ever be initialized once, and they will all get the same object returned (unless somehow a call to periodEviction snuck in the middle).

Why aren't my callback methods working properly?

I use AsyncRestTemplate to make resttemplate asynchronously.
These methods should wait all asyncresttemplate processes till done, And It will return reviewContent.
Problem is callback methods are not working, before the entire method works done. So I can't take proper return value of optionName and membershipGradeCode and reviewType should be included in reviewContent.
Could someone explain what am I missing now?
rev#1
Success callback methods change the state of reviewContent, Could it be a problem?
public ReviewContent getRepresentativeReviewContent(Long dealNo, Long categoryNo, String setId) {
Optional<Map<String, Object>> review = Optional.ofNullable(boardApi.getRepresentativeReviewContent(dealNo));
if (review.isPresent()) {
Long memberNo = Long.valueOf(review.get().get("memberNo").toString());
ReviewContent reviewContent = new ReviewContent();
ListenableFuture<ResponseEntity<Map>> optionInfo = dealApi.asyncGetDealOption(Long.valueOf(review.get().get("optionNo").toString()));
optionInfo.addCallback(success -> {
try {
reviewContent.setOptionName((String) ((Map<String, Object>) success.getBody().get("data")).get("dealTitle"));
} catch (Exception e) {
reviewContent.setOptionName(null);
}
}, failure -> LOGGER.error("asyncGetDealOption", failure.getStackTrace()));
ListenableFuture<ResponseEntity<Map>> gradeInfoOfThisMember = mktApi.asyncGetMembershipGradeOfThisMember(memberNo);
gradeInfoOfThisMember.addCallback(success -> {
try {
reviewContent.setMembershipGradeCode((Integer) ((Map<String, Object>) success.getBody().get("data")).get("grade"));
} catch (Exception e) {
reviewContent.setMembershipGradeCode(0);
}
},
failure -> {
reviewContent.setMembershipGradeCode(0);
LOGGER.error("asyncGetMembershipGradeOfThisMember", failure.getStackTrace());
});
ListenableFuture<ResponseEntity<ReviewType>> reviewTypeByCategoryNo = boardApi.asyncGetReviewTypeByCategoryNo(categoryNo, setId);
reviewTypeByCategoryNo.addCallback(success -> {
try {
reviewContent.setReviewType(success.getBody());
} catch (Exception e) {
reviewContent.setReviewType(null);
}
},
failure -> {
reviewContent.setReviewType(null);
LOGGER.error("asyncGetReviewTypeByCategoryNo", failure.getStackTrace());
});
reviewContent.setReviewCount((Integer) review.get().get("reviewCount"));
reviewContent.setReviewAvgScore((Double) review.get().get("reviewAvgScore"));
reviewContent.setContents((String) review.get().get("contents"));
reviewContent.setCreateDt((String) review.get().get("createdDt"));
reviewContent.setUpdateDt((String) review.get().get("updatedDt"));
reviewContent.setMemberSrl(memberNo);
reviewContent.setTitle((String) review.get().get("title"));
reviewContent.setAccountSrl(Long.valueOf(review.get().get("accountNo").toString()));
reviewContent.setMemberId((String) review.get().get("memberId"));
reviewContent.setAccountSrl(Long.valueOf(review.get().get("accountNo").toString()));
boolean isApiExecutionDone = false;
while (!isApiExecutionDone) {
if (gradeInfoOfThisMember.isDone() && optionInfo.isDone() && reviewTypeByCategoryNo.isDone()) {
isApiExecutionDone = true;
}
}
return reviewContent;
}
return new ReviewContent();
}
So your problem is that the callbacks set properties on the object returned by your method. However, they are also executed asynchronously, and are not part of the done status of the Future: they are themselves executed once the Future is done, concurrently with the code in the getRepresentativeReviewContent method. Since the method returns as soon as all Futures are done, the properties aren't (all) set as they should.
Moreover, you didn't show the code for your ReviewContent object, but I'm pretty sure it doesn't declare the optionType, membershipGradeCode or reviewType fields as volatile. Since there are no barriers (such as synchronized blocks or Locks) in the method, there's no guarantee in the Java Memory Model that the values set in the callbacks (i.e. in other threads) would be seen in the thread executing the getRepresentativeReviewContent method.
Callbacks should only be used for side-effects outside of your main execution path, since it's hard to coordinate with them: you would have to use things like a CountDownLatch to make sure they have all executed, that would make the code even more complex.
Just wait for the asynchronous results in a straight-forward way (the code is untested though):
try {
// Not sure why you need to catch Exception here?
// If it's for flow control (absent entry in a Map), it's a bad practice.
// Just check it instead of catching NullPointerException.
reviewContent.setOptionName((String)
((Map<String, Object>) optionInfo.get().getBody().get("data"))
.get("dealTitle"));
} catch (InterruptedException e) {
Thread.currentThread().interrupt();
LOGGER.error("asyncGetDealOption", e);
reviewContent.setOptionName(null);
} catch (CancellationException | ExecutionException e) {
LOGGER.error("asyncGetDealOption", e);
reviewContent.setOptionName(null);
}
Another option is to compose the Futures, such as what can be done with Guava's Futures.transform to actually get the string you need out of the complete response, so you can just call get() on that composed Future to set your property. You'd still have to managed the errors, though.

Timeout the thread if any bundle is taking lot of time

I am working on a project in which I will have different Bundles. Let's take an example, Suppose I have 5 bundles and each of those bundles will have a method name process.
Now currently, I am calling the process method of all those 5 bundles sequentially, one by one and then I am writing to the database. But that's what I don't want.
I need to call all those 5 Bundles process method in parallel using multithread and then write to the database.
And I also want to have some timeout feature for those threads. I will be having a default timeout settings for all the threads for the bundles. If any bundle is taking some higher time than the timeout settings I have, then I want to timeout those threads and then log back saying this bundle got timeout bcoz it was taking lot of time.
I hope question is clear enough...
Below is the code I have so far which is calling process method sequentially one by one.
public void processEvents(final Map<String, Object> eventData) {
final Map<String, String> outputs = (Map<String, String>)eventData.get(BConstants.EVENT_HOLDER);
for (final BundleRegistration.BundlesHolderEntry entry : BundleRegistration.getInstance()) {
final Map<String, String> response = entry.getPlugin().process(outputs);
// write to the database.
System.out.println(response);
}
}
I am not sure what is the best and efficient way to do this? Because, in future, it might be possible that I will have more than 5 bundles.
Can anyone provide me an example of how can I achieve this? Any help will be appreciated on this. Thanks.
It is not too difficult to achieve what you want, but you should be aware that both with concurrency and timeouts you get added complexity, especially when it comes to error handling.
For instance, threads that are running when a timeout occurs may keep running after the timeout. Only well behaved threads that cooperate by handling an interrupt signal will be able to stop succefully in the middle of processing.
You must also make sure that individual bundle entries may be processed in parallel, i.e. that the are thread safe. If they modify some shared resource while processing, then you might get strange errors as a result.
I was also wondering whether you wanted to include the database writing to each of these threads. If so, you will need to handle interruptions while writing to the database; e.g. by rolling back a transaction.
Anyways, to get thread pooling and a total timeout for all threads, you can use ExecutorService with (for instance) a fixed pool size and invoke all threads using the invokeAll method.
The following attempt is most probably flawed and error handling is by no means complete, but it should give you a starting point.
First, an implementation of Callable for your threads:
public class ProcessBundleHolderEntry implements Callable {
private BundleRegistration.BundlesHolderEntry entry;
private Map<String, String> outputs;
public ProcessBundleHolderEntry(BundleRegistration.BundlesHolderEntry entry, Map<String, String> outputs) {
this.entry = entry;
this.outputs = outputs;
}
public Object call() throws Exception {
final Map<String, String> response = entry.getPlugin().process(outputs);
// write to the database.
System.out.println(response);
return response;
}
}
and now, the modified processEvents method:
public void processEvents(final Map<String, Object> eventData) {
ExecutorService pool = Executors.newFixedThreadPool(5);
List<ProcessBundleHolderEntry> entries = new ArrayList<ProcessBundleHolderEntry>();
Map<String, String> outputs = (Map<String, String>)eventData.get(BConstants.EVENT_HOLDER);
for (BundleRegistration.BundlesHolderEntry entry : BundleRegistration.getInstance()) {
ProcessBundleHolderEntry processBundleHolderEntry = new ProcessBundleHolderEntry(entry, outputs);
entries.add(processBundleHolderEntry);
}
try {
List<Future<Object>> futures = pool.invokeAll(entries, 30, TimeUnit.SECONDS);
for (int i = 0; i < futures.size(); i++) {
// This works since the list of future objects are in the
// same sequential order as the list of entries
Future<Object> future = futures.get(i);
ProcessBundleHolderEntry entry = entries.get(i);
if (!future.isDone()) {
// log error for this entry
}
}
} catch (InterruptedException e) {
// handle this exception!
}
}
The reply given by Steinar is correct but this solution is not scalable as you said "in future, it might be possible that I will have more than 5 bundles." and I am sure that you might be adding bundles at the runtime or afterwards if some tasks are being completed and there might also be a limitation that you can execute atmost 'n' bundles parallely, in that case the executorService.InvokeAll will terminate the pending tasks which have not started if the timer specified is reached.
I have created a simple sample which might be useful to you, This example provides flexibility on how many threads you want to run parallely and you can add tasks or bundles as and when you need.
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import testproject.Bundles;
import testproject.ExecuteTimedOperation;
public class ParallelExecutor
{
public static int NUMBER_OF_PARALLEL_POLL = 4;
public static void main(String[] args)
{
ExecutorService executorService = Executors.newFixedThreadPool(NUMBER_OF_PARALLEL_POLL );
// Create bundle of objects you want
List<Bundles> lstBun = new ArrayList<Bundles>();
for (Bundles bundles : lstBun)
{
final ExecuteTimedOperation ope =new ExecuteTimedOperation(bundles, new HashMap<String, Object>());
executorService.submit(new Runnable()
{
public void run()
{
ope.ExecuteTask();
}
});
}
}
}
package testproject;
import java.util.Map;
import java.util.Random;
public class ExecuteTimedOperation
{
Bundles _bun;
Map<String, Object> _eventData;
public static long TimeInMilleToWait = 60 * 1000; //Time which each thread should wait to complete task
public ExecuteTimedOperation(Bundles bun, Map<String, Object> eventData)
{
_bun = bun;
_eventData = eventData;
}
public void ExecuteTask()
{
try
{
Thread t = new Thread(new Runnable()
{
public void run()
{
_bun.processEvents(_eventData);
}
});
t.start();
t.join(TimeInMilleToWait);
}
catch (InterruptedException e) {
//log back saying this bundle got timeout bcoz it was taking lot of time.
}
catch (Exception e) {
//All other type of exception will be handled here
}
}
}
package testproject;
import java.util.Map;
public class Bundles
{
public void processEvents(final Map<String, Object> eventData)
{
//THE code you want to execute
}
}

Calling different bundles in parallel using Multithreaded code

I am working on a project in which I will be having different Bundles. Let's take an example, Suppose I have 5 Bundles and each of those bundles will have a method name process.
Now currently, I am calling the process method of all those 5 bundles sequentially, one by one and then I am writing to the database. But that's what I don't want.
Below are the things that I am looking for-
I need to call all those 5 Bundles process method in parallel using multithreaded code and then write to the database. I am not sure what is the right way to do that? Should I have five thread? One thread for each bundle? But what will happen in that scenario, suppose if I have 50 bundles, then I will have 50 threads?
And also, I want to have timeout feature as well. If any bundles is taking lot of time than the threshold setup by us, then it should get timeout and log as an error that this bundle has taken lot of time.
I hope the question is clear enough.
Below is the code I have so far which is calling process method of each bundles sequentially one by one.
public void callBundles(final Map<String, Object> eventData) {
final Map<String, String> outputs = (Map<String, String>)eventData.get(Constants.HOLDER);
for (final BundleRegistration.BundlesHolderEntry entry : BundleRegistration.getInstance()) {
// calling the process method of a bundle
final Map<String, String> response = entry.getPlugin().process(outputs);
// then write to the database.
System.out.println(response);
}
}
I am not sure what is the best and efficient way to do this? And I don't want to write sequentially. Because, in future, it might be possible that I will have more than 5 bundles.
Can anyone provide me an example of how can I do this? I have tried doing it but somehow it is not the way I am looking for.
Any help will be appreciated on this. Thanks.
Update:-
This is what I came up with-
public void callBundles(final Map<String, Object> eventData) {
// Three threads: one thread for the database writer, five threads for the plugin processors
final ExecutorService executor = Executors.newFixedThreadPool(5);
final BlockingQueue<Map<String, String>> queue = new LinkedBlockingQueue<Map<String, String>>();
#SuppressWarnings("unchecked")
final Map<String, String> outputs = (Map<String, String>)eventData.get(Constants.EVENT_HOLDER);
for (final BundleRegistration.BundlesHolderEntry entry : BundleRegistration.getInstance()) {
executor.submit(new Runnable () {
public void run() {
final Map<String, String> response = entry.getPlugin().process(outputs);
// put the response map in the queue for the database to read
queue.offer(response);
}
});
}
Future<?> future = executor.submit(new Runnable () {
public void run() {
Map<String, String> map;
try {
while(true) {
// blocks until a map is available in the queue, or until interrupted
map = queue.take();
// write map to database
System.out.println(map);
}
} catch (InterruptedException ex) {
// IF we're catching InterruptedException then this means that future.cancel(true)
// was called, which means that the plugin processors are finished;
// process the rest of the queue and then exit
while((map = queue.poll()) != null) {
// write map to database
System.out.println(map);
}
}
}
});
// this interrupts the database thread, which sends it into its catch block
// where it processes the rest of the queue and exits
future.cancel(true); // interrupt database thread
// wait for the threads to finish
try {
executor.awaitTermination(5, TimeUnit.MINUTES);
} catch (InterruptedException e) {
//log error here
}
}
But I was not able to add any timeout feature into this yet.. And also If I am run my above code as it is, then also it is not running.. I am missing anything?
Can anybody help me with this?
This is BASIC example, partially based on the solution presented in ExecutorService that interrupts tasks after a timeout.
You will have to figure out the best way to get this implemented into your own code. Use it only as a guide!
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.Callable;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
import java.util.concurrent.Future;
import java.util.concurrent.ScheduledExecutorService;
import java.util.concurrent.TimeUnit;
public class ExecutorExample {
// This is used to "expire" long running tasks
protected static final ScheduledExecutorService EXPIRE_SERVICE = Executors.newScheduledThreadPool(1);
// This is used to manage the bundles and process them as required
protected static final ExecutorService BUNDLES_SERVICE = Executors.newFixedThreadPool(10);
public static void main(String[] args) {
// A list of the future tasks created by the BUNDLES_SERVICE.
// We need this so we can monitor the progress of the output
List<Future<String>> futureTasks = new ArrayList<>(100);
// This is a list of all the tasks that have either completed
// or begin canceled...we want these so we can determine
// the results...
List<Future<String>> completedTasks = new ArrayList<>(100);
// Add all the Bundles to the BUNDLES_SERVICE
for (int index = 0; index < 100; index++) {
Bundle bundle = new Bundle();
// We need a reference to the future so we can cancel it if we
// need to
Future<String> futureBundle = BUNDLES_SERVICE.submit(bundle);
// Set this bundles future, see Bundle for details
bundle.setFuture(futureBundle);
// Add it to our monitor queue...
futureTasks.add(futureBundle);
}
// Basically we are going to move all completed/canceled bundles
// from the "active" to the completed list and wait until there
// are no more "active" tasks
while (futureTasks.size() > 0) {
try {
// Little bit of a pressure release...
Thread.sleep(1000);
} catch (InterruptedException ex) {
}
// Check all the bundles...
for (Future<String> future : futureTasks) {
// If it has completed or was cancelled, move it to the completed
// list. AKAIK, isDone will return true is isCancelled is true as well,
// but this illustrates the point
if (future.isCancelled() || future.isDone()) {
completedTasks.add(future);
}
}
// Remove all the completed tasks from the future tasks lists
futureTasks.removeAll(completedTasks);
// Some idea of progress...
System.out.println("Still have " + futureTasks.size() + " outstanding tasks...");
}
// Dump the results...
int index = 0;
for (Future<String> future : completedTasks) {
index++;
System.out.print("Task " + index);
if (future.isCancelled()) {
System.out.println(" was canceled");
} else if (future.isDone()) {
try {
System.out.println(" completed with " + future.get());
} catch (Exception ex) {
System.out.println(" failed because of " + ex.getMessage());
}
}
}
System.exit(0);
}
public static class ExpireBundle implements Runnable {
private final Future futureBundle;
public ExpireBundle(Future futureBundle) {
this.futureBundle = futureBundle;
}
#Override
public void run() {
futureBundle.cancel(true);
}
}
public static class Bundle implements Callable<String> {
private volatile Future<String> future;
#Override
public String call() throws Exception {
// This is the tricky bit. In order to cancel a task, we
// need to wait until it runs, but we also need it's future...
// We could use another, single threaded queue to do the job
// but that's getting messy again and it won't provide the information
// we need back to the original calling thread that we are using
// to schedule and monitor the threads...
// We need to have a valid future before we can continue...
while (future == null) {
Thread.sleep(250);
}
// Schedule an expiry call for 5 seconds from NOW...this is important
// I original thought about doing this when I schedule the original
// bundle, but that precluded the fact that some tasks would not
// have started yet...
EXPIRE_SERVICE.schedule(new ExpireBundle(future), 5, TimeUnit.SECONDS);
// Sleep for a random amount of time from 1-10 seconds
Thread.sleep((long) (Math.random() * 9000) + 1000);
return "Happy";
}
protected void setFuture(Future<String> future) {
this.future = future;
}
}
}
Also. I had thought of using ExecutorService#invokeAll to wait for the tasks to complete, but this precluded the ability to timeout tasks. I don't like having to feed the Future into the Callable, but no other solution seemed to come to mind that would allow me to get the results from the submitted Future.

How to call same method of a different class in multithreaded way

I have a method named process in two of my Classes, lets say CLASS-A and CLASS-B. Now in the below loop, I am calling process method of both of my classes sequentially meaning one by one and it works fine but that is the not the way I am looking for.
for (ModuleRegistration.ModulesHolderEntry entry : ModuleRegistration.getInstance()) {
final Map<String, String> response = entry.getPlugin().process(outputs);
// write to database
System.out.println(response);
}
Is there any way, I can call the process method of both of my classes in a multithreaded way. Meaning one thread will call process method of CLASS-A and second thread will call process method of CLASS-B.
And then after that I was thinking to write the data that is being returned by the process method into the database. So I can have one more thread for writing into database.
Below is the code that I came up with in a multithreaded way but somehow it is not running at all.
public void writeEvents(final Map<String, Object> data) {
// Three threads: one thread for the database writer, two threads for the plugin processors
final ExecutorService executor = Executors.newFixedThreadPool(3);
final BlockingQueue<Map<String, String>> queue = new LinkedBlockingQueue<Map<String, String>>();
#SuppressWarnings("unchecked")
final Map<String, String> outputs = (Map<String, String>)data.get(ModelConstants.EVENT_HOLDER);
for (final ModuleRegistration.ModulesHolderEntry entry : ModuleRegistration.getInstance()) {
executor.submit(new Runnable () {
public void run() {
final Map<String, String> response = entry.getPlugin().process(outputs);
// put the response map in the queue for the database to read
queue.offer(response);
}
});
}
Future<?> future = executor.submit(new Runnable () {
public void run() {
Map<String, String> map;
try {
while(true) {
// blocks until a map is available in the queue, or until interrupted
map = queue.take();
// write map to database
System.out.println(map);
}
} catch (InterruptedException ex) {
// IF we're catching InterruptedException then this means that future.cancel(true)
// was called, which means that the plugin processors are finished;
// process the rest of the queue and then exit
while((map = queue.poll()) != null) {
// write map to database
System.out.println(map);
}
}
}
});
// this interrupts the database thread, which sends it into its catch block
// where it processes the rest of the queue and exits
future.cancel(true); // interrupt database thread
// wait for the threads to finish
try {
executor.awaitTermination(5, TimeUnit.MINUTES);
} catch (InterruptedException e) {
//log error here
}
}
But If I remove the last line executor.awaitTermination(5, TimeUnit.MINUTES); then it start running fine and after some time, I always get error like this-
JVMDUMP006I Processing dump event "systhrow", detail "java/lang/OutOfMemoryError" - please wait.
JVMDUMP032I JVM requested Heap dump using 'S:\GitViews\Stream\goldseye\heapdump.20130827.142415.16456.0001.phd' in response to an event
JVMDUMP010I Heap dump written to S:\GitViews\Stream\goldseye\heapdump.20130827.142415.16456.0001.phd
JVMDUMP006I Processing dump event "systhrow", detail "java/lang/OutOfMemoryError" - please wait.
Can anybody help me in figuring out what's the problem and what wrong I am doing in my above code? if I am running sequentially then I don't get any errors and it works fine.
And also is there any better way of doing this as compared to the way I am doing? Because in future I can have multiple plugin processor as compared to two.
What I am trying to do is- Call the process method of both of my classes in a multithreaded way and then write into the database bcoz my process method will return back a Map.
Any help will be appreciated on this.. And I am looking for a workable example on this if possible. Thanks for the help,
The code snippet you pasted has few issues, if you fix them, this should work.
1. You are using an infinite loop to fetch element from the blocking queue and trying to break this using future. This is definitely not a good approach. The problem with this approach is it is possible that your database thread would never run because it could be cancelled by the future task running in the caller thread even before it runs. This is error-prone.
- You should run the while loop fixed number of times (you already know how many producers are there or how many times you are going to get the response).
Also, tasks submitted to executor service should be independent tasks...here your database task is dependent on the execution of other tasks..this can also lead to deadlock if your execution policy changes..for example if you use single thread pool executor and if database thread is scheduled it would just block waiting for producers to add data in the queue.
A good way is to create task that retrieves data and update the database in the same thread.
Or retrieve all the responses first and then execute database operations in parallel
public void writeEvents(final Map data) {
final ExecutorService executor = Executors.newFixedThreadPool(3);
#SuppressWarnings("unchecked")
final Map<String, String> outputs = (Map<String, String>)data.get(ModelConstants.EVENT_HOLDER);
for (final ModuleRegistration.ModulesHolderEntry entry : ModuleRegistration.getInstance()) {
executor.submit(new Runnable () {
public void run() {
try {
final Map<String, String> response = entry.getPlugin().process(outputs);
//process the response and update database.
System.out.println(map);
} catch (Throwable e) {
//handle execption
} finally {
//clean up resources
}
}
});
}
// This will wait for running threads to complete ..it's an orderly shutdown.
executor.shutdown();
}
OK, here's some code for the comments I suggested above. Disclaimer: I'm not sure whether it works or even compiles, or whether it solves the problem. But the idea is to take control of the cancellation process instead of relying on future.cancel which I suspect could cause problems.
class CheckQueue implements Runnable {
private volatile boolean cancelled = false;
public void cancel() { cancelled = true; }
public void run() {
Map<String, String> map;
try {
while(!cancelled) {
// blocks until a map is available in the queue, or until interrupted
map = queue.take();
if (cancelled) break;
// write map to database
System.out.println(map);
} catch (InterruptedException e) {
}
while((map = queue.poll()) != null) {
// write map to database
System.out.println(map);
}
}
}
CheckQueue queueChecker = new CheckQueue ();
Future<?> future = executor.submit(queueChecker);
// this interrupts the database thread, which sends it into its catch block
// where it processes the rest of the queue and exits
queueChecker.cancel();

Categories