Java fork join issue - concurrency

I am learning fork-join technique in java and have written the following program. I am running a for loop (5 times) and I want to run the contents of the for loop in separate threads. This is happening correctly. The problem is that when all the threads are finished, I want a vector of size 5 and it must contain the result of execution of thread 1 at index 0, result of execution of thread 2 at index 1 ............ result of execution of thread 5 at index 4. I am cleanly visualize what I want to achieve but don't know the syntax for it.
Currently I just get 1 number in my vector.
import java.util.Random;
import java.util.Vector;
import java.util.concurrent.*;
public class App {
public static void main(String[] args) {
// TODO Auto-generated method stub
ExecutorService executor = Executors.newCachedThreadPool();
Future<Vector<Integer> > futureResult = null;
for(int i = 0; i < 5; i++){
futureResult = executor.submit(new Callable<Vector<Integer> >(){
#Override
public Vector<Integer> call() throws Exception {
Random random = new Random();
int duration = random.nextInt(4000);
Vector<Integer> v = new Vector<Integer>();
v.add(duration);
return v;
}
});
}
executor.shutdown();
try {
System.out.println(futureResult.get());
} catch (InterruptedException | ExecutionException e) {
e.printStackTrace();
}
}
}

You're creating a new vector in every thread, without any connection with the shared memory. You could just pass the index to write on to every thread, and also change the vector to an array or something pre-allocated.
Another option would be to allocate the 5 vector positions and send every Integer object to its thread.

Related

Why am I getting a "read access violation" exception thrown while accessing value from std::future?

Edit: my question is different from the suggested question becaude I cannot poll 100's of future.get() and wait to execute the remaining program in main() thread. I am hoping the threads to return the value when the thread is done executing and then the calling function should run.
As per this question
in C++ future waits for the results and halts the next line of code
I am having the following code where I am creating about 100 async threads and then retrieving the return variable;
void BEVTest::runTests()
{
CameraToBEV cbevObj(RoiBbox, OutputDim, InputDim);
// for multithreading
std::vector<std::future<cv::Mat>> processingThread;
std::vector<cv::Mat> opMatArr;
for (int i = 0; i < jsonObjects.size(); i++) {
processingThread.emplace_back(std::async(std::launch::async, &CameraToBEV::process,&cbevObj, std::ref(jsonObjects[i]), RoiBbox, OutputDim, InputDim));
}
for (auto& future : processingThread) {
opMatArr.emplace_back(future.get());
}
}
I am getting a run time exception of "read access violation" at the line opMatArr.emplace_back(future.get());
when I checked the processingThread variable, it shows all the future values as pending. So, if the above quote from the question is correct, shouldn't my code wait till it gets all the future values? Else otherwise this answer provides the following solution to wait till you get the future value from future.get()
auto f = std::async(std::launch::async, my_func)
.then([] (auto fut) {
auto result = fut.get();
/* Do stuff when result is ready. */
});
But this won't be possible for me because then I will have to poll all 100 future.get() and it will have an overhead which will defeat the purpose of creating threads.
I have two questions;
1> Why am I getting a run time exception?
2> How do I wait till all 100's of future.get() return value?
EDIT: I am providing my process function.
cv::Mat CameraToBEV::process(json& jsonObject, std::vector<Point2f> roibbox, Point2f opDim, Point2f ipDim)
{ // Parameters
img_width = imgWidth = opDim.x;
img_height = imgHeight = opDim.y;
for (int i=0; i<roibbox.size(); i++)
{
roiPoints[i] = roibbox[i];
}
// From Centroids class
jsonObj = jsonObject;
initializeCentroidMatrix();
getBBoxes();
transformCoordinates();
// From Plotter class
cv::Mat resultImg = plotLocation(finalCentroids,violationArr, ipDim, opDim);
return resultImg;
}

Simple program with Callable Future never terminates

I was playing around with Callable and Future and stumbled upon an issue.
This is a piece of code that never terminates and times out even though the IDE allows 5 seconds to run, and the code does not need more than 3 seconds (It gives a Time Limit Exceeded error): https://ideone.com/NcL0YV
/* package whatever; // don't place package name! */
import java.util.*;
import java.util.concurrent.*;
import java.lang.*;
import java.io.*;
/* Name of the class has to be "Main" only if the class is public. */
class Ideone
{
public static void main (String[] args) throws java.lang.Exception
{
Ideone obj = new Ideone();
Future<Integer> res = obj.doCallable();
System.out.println(res.get());
}
public Future<Integer> calculate(Integer input) {
ExecutorService executor = Executors.newFixedThreadPool(1);
return executor.submit(() -> {
long start = System.currentTimeMillis();
Thread.sleep(2000);
System.out.println("Sleep time in ms = "+(System.currentTimeMillis()-start));
return input * input;
});
}
public Future<Integer> doCallable() {
int value = 99;
try {
Callable<Future> callable = () -> calculate(value);
Future<Integer> future = callable.call();
return future;
} catch (final Exception e) {
e.printStackTrace();
throw new RuntimeException(e);
}
}
}
This is a similar piece of code that terminates because I added "System.exit(0)" (which is not advisable): https://ideone.com/HDvl7y
/* package whatever; // don't place package name! */
import java.util.*;
import java.util.concurrent.*;
import java.lang.*;
import java.io.*;
/* Name of the class has to be "Main" only if the class is public. */
class Ideone
{
public static void main (String[] args) throws java.lang.Exception
{
Ideone obj = new Ideone();
Future<Integer> res = obj.doCallable();
System.out.println(res.get());
System.exit(0);
}
public Future<Integer> calculate(Integer input) {
ExecutorService executor = Executors.newFixedThreadPool(1);
return executor.submit(() -> {
long start = System.currentTimeMillis();
Thread.sleep(2000);
System.out.println("Sleep time in ms = "+(System.currentTimeMillis()-start));
return input * input;
});
}
public Future<Integer> doCallable() {
int value = 99;
try {
Callable<Future> callable = () -> calculate(value);
Future<Integer> future = callable.call();
return future;
} catch (final Exception e) {
e.printStackTrace();
throw new RuntimeException(e);
}
}
}
Please help me understand why do we need System.exit(0) or shutdown() even if callable task is complete (future.get() call is a blocking call).
EDIT:
I did all the above to solve the main issue of ever-increasing threads in my application because of the below code snippet. I am not sure how to complete this future automatically after a certain timeout without involving the main thread (which exits right away).
#Override
public void publish(#NonNull final String message,
#NonNull final String topicArn) throws PublishingException {
if (!publishAsync(message, topicArn)) {
throw new PublishingException("Publish attempt failed for the message:"
+ message);
}
}
private boolean publishAsync(final String message,
final String topicArn) {
Callable<Future> publishCallable = () -> snsClient.publishAsync(topicArn, message);
try {
Future<PublishResult> result = publishCallable.call();
log.debug("Asynchronously published message {} to SNS topic {}.", message, topicArn);
return !result.isDone() || result.get().getMessageId() != null;
} catch (final Exception e) {
return false;
}
}
void shutdown()
Initiates an orderly shutdown in which previously submitted tasks are executed, but no new tasks will be accepted. Invocation has no additional effect if already shut down.
This method does not wait for previously submitted tasks to complete execution. Use awaitTerminationto do that.

c++11 multi-reader / multi-writer queue using atomics for object state and perpetual incremented indexes

I am using atomics and a circular buffer in order to implement a multi-reader threads, multi-writer threads object pool.
It is difficult to investigate because instrumenting code leads to bug vanishment !
The model
Producers (or writer threads) request an Element to the Ring in order to 'prepare' the element. When terminated, the writer thread changes the element state so a reader can 'consume' it. After that, the element becomes available again for writing.
Consumers (or reader threads) request an object to the Ring in order to 'read' the object.
After 'releasing' the object, the object is in a state::Ready state, eg available to be consume by a reader thread.
It can fail if no object is available eg the next free object in the Ring is not on state::Unused state.
The 2 classes, Element and Ring
Element :
to be written, a writer thread must successfully exchange the _state member from state::Unused to state::LockForWrite
when finished, the writer thread force the state to state::Ready (it should be the only to handle this Element)
to be read, a rader thread must successfully exchange the _state member from state::Ready to state::LockForRead
when finished, the reader thread force the state to state::Unused (it should be the only to handle this Element)
Summarized :
writers lifecycle : state::Unused -> state::LockForWrite -> state::Ready
readers lifecycle : state::Ready -> state::LockForRead -> state::Unused
Ring
has a vector of Element , seen as a circular buffer.
std::atomic<int64_t> _read, _write; are the 2 indexes used to access the elements via :
_elems[ _write % _elems.size() ] for writers,
_elems[ _read % _elems.size() ] for readers.
When a reader has successfully LockForRead an object, the _read index is incremented.
When a writer has successfully LockForWrite an object, the _write index is incremented.
The main :
We add to a vector some writers and readers threads sharing the same Ring. Each thread just try to get_read or get_write element and release them just after.
Based on Element transition everything should be fine but one can observe that the Ring at some point gets blocked like because some elements in the ring are in state state::Ready with a _write % _elems.size() index pointing on it and symetrically, some elements in the ring are in state state::Unused with a _read % _elems.size() index pointing on it ! Both = deadlock.
#include<atomic>
#include<vector>
#include<thread>
#include<iostream>
#include<cstdint>
typedef enum : int
{
Unused, LockForWrite, Ready, LockForRead
}state;
class Element
{
std::atomic<state> _state;
public:
Element():_state(Unused){ }
// a reader need to successfully make the transition Ready => LockForRead
bool lock_for_read() { state s = Ready; return _state.compare_exchange_strong(s, LockForRead); }
void unlock_read() { state s = Unused; _state.store(s); }
// a reader need to successfully make the transition Unused => LockForWrite
bool lock_for_write() { state s = Unused; return _state.compare_exchange_strong(s, LockForWrite); }
void unlock_write() { state s = Ready; _state.store(s); }
};
class Ring
{
std::vector<Element> _elems;
std::atomic<int64_t> _read, _write;
public:
Ring(size_t capacity)
: _elems(capacity), _read(0), _write(0) {}
Element * get_for_read() {
Element * ret = &_elems[ _read.load() % _elems.size() ];
if (!ret->lock_for_read()) // if success, the object belongs to the caller thread as reader
return NULL;
_read.fetch_add(1); // success! incr _read index
return ret;
}
Element * get_for_write() {
Element * ret = &_elems[ _write.load() % _elems.size() ];
if (!ret->lock_for_write())// if success, the object belongs to the caller thread as writer
return NULL;
_write.fetch_add(1); // success! incr _write index
return ret;
}
void release_read(Element* e) { e->unlock_read();}
void release_write(Element* e) { e->unlock_write();}
};
int main()
{
const int capacity = 10; // easy to process modulo[![enter image description here][1]][1]
std::atomic<bool> stop=false;
Ring ring(capacity);
std::function<void()> writer_job = [&]()
{
std::cout << "writer starting" << std::endl;
Element * e;
while (!stop)
{
if (!(e = ring.get_for_write()))
continue;
// do some real writer job ...
ring.release_write(e);
}
};
std::function<void()> reader_job = [&]()
{
std::cout << "reader starting" << std::endl;
Element * e;
while (!stop)
{
if (!(e = ring.get_for_read()))
continue;
// do some real reader job ...
ring.release_read(e);
}
};
int nb_writers = 1;
int nb_readers = 2;
std::vector<std::thread> threads;
threads.reserve(nb_writers + nb_readers);
std::cout << "adding writers" << std::endl;
while (nb_writers--)
threads.push_back(std::thread(writer_job));
std::cout << "adding readers" << std::endl;
while (nb_readers--)
threads.push_back(std::thread(reader_job));
// wait user key press, halt in debugger after 1 or 2 seconds
// in order to reproduce problem and watch ring
std::cin.get();
stop = true;
std::cout << "waiting all threads...\n";
for (auto & th : threads)
th.join();
std::cout << "end" << std::endl;
}
This "watch debugger screeshot" has been took pausing the program after running 1 second. As you can see, _read is pointing to the element 8 marked as state::Unused so no transition can unblock this state for this reader, except a writer but _write index is pointing on element 0 with state state::Ready !
My question: what did I missed in this ? Structurally I am sure the sequence is correct but I am missing some atomic trick ...
os tested : rhel5/gcc 4.1.2, rhel 7/gcc 4.8, win10/ms visual 2015, win10/mingw
Yann's answer is correct about the problem: your threads can create "holes" in the sequence by reading and writing elements out-of-order if there's a delay between the read/write lock and the increment of the index. The fix is to verify that the index has not changed between the initial read and the increment, a la:
class Element
{
std::atomic<state> _state;
public:
Element():_state(Unused){ }
// a reader need to successfully make the transition Ready => LockForRead
bool lock_for_read() {
state s = Ready;
return _state.compare_exchange_strong(s, LockForRead);
}
void abort_read() { _state = Ready; }
void unlock_read() { state s = Unused; _state.store(s); }
// a reader need to successfully make the transition Unused => LockForWrite
bool lock_for_write() {
state s = Unused;
return _state.compare_exchange_strong(s, LockForWrite);
}
void abort_write() { _state = Unused; }
void unlock_write() { state s = Ready; _state.store(s); }
};
class Ring
{
std::vector<Element> _elems;
std::atomic<int64_t> _read, _write;
public:
Ring(size_t capacity)
: _elems(capacity), _read(0), _write(0) {}
Element * get_for_read() {
auto i = _read.load();
Element * ret = &_elems[ i % _elems.size() ];
if (ret->lock_for_read()) {
// if success, the object belongs to the caller thread as reader
if (_read.compare_exchange_strong(i, i + 1))
return ret;
// Woops, reading out of order.
ret->abort_read();
}
return NULL;
}
Element * get_for_write() {
auto i = _write.load();
Element * ret = &_elems[ i % _elems.size() ];
if (ret->lock_for_write()) {
// if success, the object belongs to the caller thread as writer
if (_write.compare_exchange_strong(i, i + 1))
return ret;
// Woops, writing out of order.
ret->abort_write();
}
return NULL;
}
void release_read(Element* e) { e->unlock_read();}
void release_write(Element* e) { e->unlock_write();}
};
You do not have atomic section around the increment of the two shared counters _read and _write.
That looks bad to me, you could switch another element without meaning to.
Imagine this scenario,
1 reader R1 and 1 writer W are happily cooperating.
Reader 2 executes : Element * ret = &_elems[ _read.load() % _elems.size() ];
and gets pushed off the cpu.
Now R1 and W are still playing together, so the positions of _read and _write are now arbitrary w.r.t. the element ret that R2 is pointing.
Now at some point R2 gets scheduled, and it so happens that *ret_ is readable (again possibly, R1 and W went around the block a few times).
Ouch, as you see, we will read it, and increment "_read", but _read has no relation to _ret. This creates kind of holes, elements that have not been read, but that are below _read index.
So, make critical sections to ensure that increment of _read/_write is done in the same semantic step as the actual lock.

D parallel loop

First, how D create parallel foreach (underlying logic)?
int main(string[] args)
{
int[] arr;
arr.length = 100000000;
/* Why it is working?, it's simple foreach which working with
reference to int from arr, parallel function return ParallelForeach!R
(ParallelForeach!int[]), but I don't know what it is.
Parallel function is part od phobos library, not D builtin function, then what
kind of magic is used for this? */
foreach (ref e;parallel(arr))
{
e = 100;
}
foreach (ref e;parallel(arr))
{
e *= e;
}
return 0;
}
And second, why it is slower then simple foreach?
Finally, If I create my own taskPool (and don't use global taskPool object), program never end. Why?
parallel returns a struct (of type ParallelForeach) that implements the opApply(int delegate(...)) foreach overload.
when called the struct submits a parallel function to the private submitAndExecute which submits the same task to all threads in the pool.
this then does:
scope(failure)
{
// If an exception is thrown, all threads should bail.
atomicStore(shouldContinue, false);
}
while (atomicLoad(shouldContinue))
{
immutable myUnitIndex = atomicOp!"+="(workUnitIndex, 1);
immutable start = workUnitSize * myUnitIndex;
if(start >= len)
{
atomicStore(shouldContinue, false);
break;
}
immutable end = min(len, start + workUnitSize);
foreach(i; start..end)
{
static if(withIndex)
{
if(dg(i, range[i])) foreachErr();
}
else
{
if(dg(range[i])) foreachErr();
}
}
}
where workUnitIndex and shouldContinue are shared variables and dg is the foreach delegate
The reason it is slower is simply because of the overhead required to pass the function to the threads in the pool and atomically accessing the shared variables.
the reason your custom pool doesn't shut down is likely you don't shut down the threadpool with finish

Good example of livelock?

I understand what livelock is, but I was wondering if anyone had a good code-based example of it? And by code-based, I do not mean "two people trying to get past each other in a corridor". If I read that again, I'll lose my lunch.
Here's a very simple Java example of livelock where a husband and wife are trying to eat soup, but only have one spoon between them. Each spouse is too polite, and will pass the spoon if the other has not yet eaten.
public class Livelock {
static class Spoon {
private Diner owner;
public Spoon(Diner d) { owner = d; }
public Diner getOwner() { return owner; }
public synchronized void setOwner(Diner d) { owner = d; }
public synchronized void use() {
System.out.printf("%s has eaten!", owner.name);
}
}
static class Diner {
private String name;
private boolean isHungry;
public Diner(String n) { name = n; isHungry = true; }
public String getName() { return name; }
public boolean isHungry() { return isHungry; }
public void eatWith(Spoon spoon, Diner spouse) {
while (isHungry) {
// Don't have the spoon, so wait patiently for spouse.
if (spoon.owner != this) {
try { Thread.sleep(1); }
catch(InterruptedException e) { continue; }
continue;
}
// If spouse is hungry, insist upon passing the spoon.
if (spouse.isHungry()) {
System.out.printf(
"%s: You eat first my darling %s!%n",
name, spouse.getName());
spoon.setOwner(spouse);
continue;
}
// Spouse wasn't hungry, so finally eat
spoon.use();
isHungry = false;
System.out.printf(
"%s: I am stuffed, my darling %s!%n",
name, spouse.getName());
spoon.setOwner(spouse);
}
}
}
public static void main(String[] args) {
final Diner husband = new Diner("Bob");
final Diner wife = new Diner("Alice");
final Spoon s = new Spoon(husband);
new Thread(new Runnable() {
public void run() { husband.eatWith(s, wife); }
}).start();
new Thread(new Runnable() {
public void run() { wife.eatWith(s, husband); }
}).start();
}
}
Run the program and you'll get:
Bob: You eat first my darling Alice!
Alice: You eat first my darling Bob!
Bob: You eat first my darling Alice!
Alice: You eat first my darling Bob!
Bob: You eat first my darling Alice!
Alice: You eat first my darling Bob!
...
This will go on forever if uninterrupted. This is a livelock because both Alice and Bob are repeatedly asking each other to go first in an infinite loop (hence live). In a deadlock situation, both Alice and Bob would simply be frozen waiting on each other to go first — they won't be doing anything except wait (hence dead).
Flippant comments aside, one example which is known to come up is in code which tries to detect and handle deadlock situations. If two threads detect a deadlock, and try to "step aside" for each other, without care they will end up being stuck in a loop always "stepping aside" and never managing to move forwards.
By "step aside" I mean that they would release the lock and attempt to let the other one acquire it. We might imagine the situation with two threads doing this (pseudocode):
// thread 1
getLocks12(lock1, lock2)
{
lock1.lock();
while (lock2.locked())
{
// attempt to step aside for the other thread
lock1.unlock();
wait();
lock1.lock();
}
lock2.lock();
}
// thread 2
getLocks21(lock1, lock2)
{
lock2.lock();
while (lock1.locked())
{
// attempt to step aside for the other thread
lock2.unlock();
wait();
lock2.lock();
}
lock1.lock();
}
Race conditions aside, what we have here is a situation where both threads, if they enter at the same time will end up running in the inner loop without proceeding. Obviously this is a simplified example. A naiive fix would be to put some kind of randomness in the amount of time the threads would wait.
The proper fix is to always respect the lock heirarchy. Pick an order in which you acquire the locks and stick to that. For example if both threads always acquire lock1 before lock2, then there is no possibility of deadlock.
As there is no answer marked as accepted answer, I have attempted to create live lock example;
Original program was written by me in Apr 2012 to learn various concept of multithreading. This time I have modified it to create deadlock, race condition, livelock etc.
So let's understand the problem statement first;
Cookie Maker Problem
There are some ingredient containers: ChocoPowederContainer, WheatPowderContainer. CookieMaker takes some amount of powder from ingredient containers to bake a Cookie. If a cookie maker finds a container empty it checks for another container to save time. And waits until Filler fills the required container. There is a Filler who checks container on regular interval and fills some quantity if a container needs it.
Please check the complete code on github;
Let me explain you implementation in brief.
I start Filler as daemon thread. So it'll keep filling containers on regular interval. To fill a container first it locks the container -> check if it needs some powder -> fills it -> signal all makers who are waiting for it -> unlock container.
I create CookieMaker and set that it can bake up to 8 cookies in parallel. And I start 8 threads to bake cookies.
Each maker thread creates 2 callable sub-thread to take powder from containers.
sub-thread takes a lock on a container and check if it has enough powder. If not, wait for some time. Once Filler fills the container, it takes the powder, and unlock the container.
Now it completes other activities like: making mixture and baking etc.
Let's have a look in the code:
CookieMaker.java
private Integer getMaterial(final Ingredient ingredient) throws Exception{
:
container.lock();
while (!container.getIngredient(quantity)) {
container.empty.await(1000, TimeUnit.MILLISECONDS);
//Thread.sleep(500); //For deadlock
}
container.unlock();
:
}
IngredientContainer.java
public boolean getIngredient(int n) throws Exception {
:
lock();
if (quantityHeld >= n) {
TimeUnit.SECONDS.sleep(2);
quantityHeld -= n;
unlock();
return true;
}
unlock();
return false;
}
Everything runs fine until Filler is filling the containers. But if I forget to start the filler, or filler goes on unexpected leave, sub-threads keep changing their states to allow other maker to go and check the container.
I have also create a daemon ThreadTracer which keeps watch on thread states and deadlocks. This the output from console;
2016-09-12 21:31:45.065 :: [Maker_0:WAITING, Maker_1:WAITING, Maker_2:WAITING, Maker_3:WAITING, Maker_4:WAITING, Maker_5:WAITING, Maker_6:WAITING, Maker_7:WAITING, pool-7-thread-1:TIMED_WAITING, pool-7-thread-2:TIMED_WAITING, pool-8-thread-1:TIMED_WAITING, pool-8-thread-2:TIMED_WAITING, pool-6-thread-1:TIMED_WAITING, pool-6-thread-2:TIMED_WAITING, pool-5-thread-1:TIMED_WAITING, pool-5-thread-2:TIMED_WAITING, pool-1-thread-1:TIMED_WAITING, pool-3-thread-1:TIMED_WAITING, pool-2-thread-1:TIMED_WAITING, pool-1-thread-2:TIMED_WAITING, pool-4-thread-1:TIMED_WAITING, pool-4-thread-2:RUNNABLE, pool-3-thread-2:TIMED_WAITING, pool-2-thread-2:TIMED_WAITING]
2016-09-12 21:31:45.065 :: [Maker_0:WAITING, Maker_1:WAITING, Maker_2:WAITING, Maker_3:WAITING, Maker_4:WAITING, Maker_5:WAITING, Maker_6:WAITING, Maker_7:WAITING, pool-7-thread-1:TIMED_WAITING, pool-7-thread-2:TIMED_WAITING, pool-8-thread-1:TIMED_WAITING, pool-8-thread-2:TIMED_WAITING, pool-6-thread-1:TIMED_WAITING, pool-6-thread-2:TIMED_WAITING, pool-5-thread-1:TIMED_WAITING, pool-5-thread-2:TIMED_WAITING, pool-1-thread-1:TIMED_WAITING, pool-3-thread-1:TIMED_WAITING, pool-2-thread-1:TIMED_WAITING, pool-1-thread-2:TIMED_WAITING, pool-4-thread-1:TIMED_WAITING, pool-4-thread-2:TIMED_WAITING, pool-3-thread-2:TIMED_WAITING, pool-2-thread-2:TIMED_WAITING]
WheatPowder Container has 0 only.
2016-09-12 21:31:45.082 :: [Maker_0:WAITING, Maker_1:WAITING, Maker_2:WAITING, Maker_3:WAITING, Maker_4:WAITING, Maker_5:WAITING, Maker_6:WAITING, Maker_7:WAITING, pool-7-thread-1:TIMED_WAITING, pool-7-thread-2:TIMED_WAITING, pool-8-thread-1:TIMED_WAITING, pool-8-thread-2:TIMED_WAITING, pool-6-thread-1:TIMED_WAITING, pool-6-thread-2:TIMED_WAITING, pool-5-thread-1:TIMED_WAITING, pool-5-thread-2:TIMED_WAITING, pool-1-thread-1:TIMED_WAITING, pool-3-thread-1:TIMED_WAITING, pool-2-thread-1:TIMED_WAITING, pool-1-thread-2:TIMED_WAITING, pool-4-thread-1:TIMED_WAITING, pool-4-thread-2:TIMED_WAITING, pool-3-thread-2:TIMED_WAITING, pool-2-thread-2:RUNNABLE]
2016-09-12 21:31:45.082 :: [Maker_0:WAITING, Maker_1:WAITING, Maker_2:WAITING, Maker_3:WAITING, Maker_4:WAITING, Maker_5:WAITING, Maker_6:WAITING, Maker_7:WAITING, pool-7-thread-1:TIMED_WAITING, pool-7-thread-2:TIMED_WAITING, pool-8-thread-1:TIMED_WAITING, pool-8-thread-2:TIMED_WAITING, pool-6-thread-1:TIMED_WAITING, pool-6-thread-2:TIMED_WAITING, pool-5-thread-1:TIMED_WAITING, pool-5-thread-2:TIMED_WAITING, pool-1-thread-1:TIMED_WAITING, pool-3-thread-1:TIMED_WAITING, pool-2-thread-1:TIMED_WAITING, pool-1-thread-2:TIMED_WAITING, pool-4-thread-1:TIMED_WAITING, pool-4-thread-2:TIMED_WAITING, pool-3-thread-2:TIMED_WAITING, pool-2-thread-2:TIMED_WAITING]
You'll notice that sub-threads and changing their states and waiting.
A real (albeit without exact code) example is two competing processes live locking in an attempt to correct for a SQL server deadlock, with each process using the same wait-retry algorithm for retrying. While it's the luck of timing, I have seen this happen on separate machines with similar performance characteristics in response to a message added to an EMS topic (e.g. saving an update of a single object graph more than once), and not being able to control the lock order.
A good solution in this case would be to have competing consumers (prevent duplicate processing as high up in the chain as possible by partitioning the work on unrelated objects).
A less desirable (ok, dirty-hack) solution is to break the timing bad luck (kind of force differences in processing) in advance or break it after deadlock by using different algorithms or some element of randomness. This could still have issues because its possible the lock taking order is "sticky" for each process, and this takes a certain minimum of time not accounted for in the wait-retry.
Yet another solution (at least for SQL Server) is to try a different isolation level (e.g. snapshot).
I coded up the example of 2 persons passing in a corridor. The two threads will avoid each other as soon as they realise their directions are the same.
public class LiveLock {
public static void main(String[] args) throws InterruptedException {
Object left = new Object();
Object right = new Object();
Pedestrian one = new Pedestrian(left, right, 0); //one's left is one's left
Pedestrian two = new Pedestrian(right, left, 1); //one's left is two's right, so have to swap order
one.setOther(two);
two.setOther(one);
one.start();
two.start();
}
}
class Pedestrian extends Thread {
private Object l;
private Object r;
private Pedestrian other;
private Object current;
Pedestrian (Object left, Object right, int firstDirection) {
l = left;
r = right;
if (firstDirection==0) {
current = l;
}
else {
current = r;
}
}
void setOther(Pedestrian otherP) {
other = otherP;
}
Object getDirection() {
return current;
}
Object getOppositeDirection() {
if (current.equals(l)) {
return r;
}
else {
return l;
}
}
void switchDirection() throws InterruptedException {
Thread.sleep(100);
current = getOppositeDirection();
System.out.println(Thread.currentThread().getName() + " is stepping aside.");
}
public void run() {
while (getDirection().equals(other.getDirection())) {
try {
switchDirection();
Thread.sleep(100);
} catch (InterruptedException e) {}
}
}
}
C# version of jelbourn's code:
using System;
using System.Runtime.CompilerServices;
using System.Threading;
using System.Threading.Tasks;
namespace LiveLockExample
{
static class Program
{
public static void Main(string[] args)
{
var husband = new Diner("Bob");
var wife = new Diner("Alice");
var s = new Spoon(husband);
Task.WaitAll(
Task.Run(() => husband.EatWith(s, wife)),
Task.Run(() => wife.EatWith(s, husband))
);
}
public class Spoon
{
public Spoon(Diner diner)
{
Owner = diner;
}
public Diner Owner { get; private set; }
[MethodImpl(MethodImplOptions.Synchronized)]
public void SetOwner(Diner d) { Owner = d; }
[MethodImpl(MethodImplOptions.Synchronized)]
public void Use()
{
Console.WriteLine("{0} has eaten!", Owner.Name);
}
}
public class Diner
{
public Diner(string n)
{
Name = n;
IsHungry = true;
}
public string Name { get; private set; }
private bool IsHungry { get; set; }
public void EatWith(Spoon spoon, Diner spouse)
{
while (IsHungry)
{
// Don't have the spoon, so wait patiently for spouse.
if (spoon.Owner != this)
{
try
{
Thread.Sleep(1);
}
catch (ThreadInterruptedException e)
{
}
continue;
}
// If spouse is hungry, insist upon passing the spoon.
if (spouse.IsHungry)
{
Console.WriteLine("{0}: You eat first my darling {1}!", Name, spouse.Name);
spoon.SetOwner(spouse);
continue;
}
// Spouse wasn't hungry, so finally eat
spoon.Use();
IsHungry = false;
Console.WriteLine("{0}: I am stuffed, my darling {1}!", Name, spouse.Name);
spoon.SetOwner(spouse);
}
}
}
}
}
Consider a UNIX system having 50 process slots.
Ten programs are running, each of which having to create 6 (sub)processes.
After each process has created 4 processes, the 10 original processes and the 40 new processes have exhausted the table. Each of the 10 original processes now sits in an endless loop forking and failing – which is aptly the situation of a livelock. The probability of this happening is very little but it could happen.
One example here might be using a timed tryLock to obtain more than one lock and if you can't obtain them all, back off and try again.
boolean tryLockAll(Collection<Lock> locks) {
boolean grabbedAllLocks = false;
for(int i=0; i<locks.size(); i++) {
Lock lock = locks.get(i);
if(!lock.tryLock(5, TimeUnit.SECONDS)) {
grabbedAllLocks = false;
// undo the locks I already took in reverse order
for(int j=i-1; j >= 0; j--) {
lock.unlock();
}
}
}
}
I could imagine such code would be problematic as you have lots of threads colliding and waiting to obtain a set of locks. But I'm not sure this is very compelling to me as a simple example.
Python version of jelbourn's code:
import threading
import time
lock = threading.Lock()
class Spoon:
def __init__(self, diner):
self.owner = diner
def setOwner(self, diner):
with lock:
self.owner = diner
def use(self):
with lock:
"{0} has eaten".format(self.owner)
class Diner:
def __init__(self, name):
self.name = name
self.hungry = True
def eatsWith(self, spoon, spouse):
while(self.hungry):
if self != spoon.owner:
time.sleep(1) # blocks thread, not process
continue
if spouse.hungry:
print "{0}: you eat first, {1}".format(self.name, spouse.name)
spoon.setOwner(spouse)
continue
# Spouse was not hungry, eat
spoon.use()
print "{0}: I'm stuffed, {1}".format(self.name, spouse.name)
spoon.setOwner(spouse)
def main():
husband = Diner("Bob")
wife = Diner("Alice")
spoon = Spoon(husband)
t0 = threading.Thread(target=husband.eatsWith, args=(spoon, wife))
t1 = threading.Thread(target=wife.eatsWith, args=(spoon, husband))
t0.start()
t1.start()
t0.join()
t1.join()
if __name__ == "__main__":
main()
I modify the answer of #jelbourn.
When one of them notices that the other is hungry, he(her) should release the spoon and wait another notify, so a livelock happens.
public class LiveLock {
static class Spoon {
Diner owner;
public String getOwnerName() {
return owner.getName();
}
public void setOwner(Diner diner) {
this.owner = diner;
}
public Spoon(Diner diner) {
this.owner = diner;
}
public void use() {
System.out.println(owner.getName() + " use this spoon and finish eat.");
}
}
static class Diner {
public Diner(boolean isHungry, String name) {
this.isHungry = isHungry;
this.name = name;
}
private boolean isHungry;
private String name;
public String getName() {
return name;
}
public void eatWith(Diner spouse, Spoon sharedSpoon) {
try {
synchronized (sharedSpoon) {
while (isHungry) {
while (!sharedSpoon.getOwnerName().equals(name)) {
sharedSpoon.wait();
//System.out.println("sharedSpoon belongs to" + sharedSpoon.getOwnerName())
}
if (spouse.isHungry) {
System.out.println(spouse.getName() + "is hungry,I should give it to him(her).");
sharedSpoon.setOwner(spouse);
sharedSpoon.notifyAll();
} else {
sharedSpoon.use();
sharedSpoon.setOwner(spouse);
isHungry = false;
}
Thread.sleep(500);
}
}
} catch (InterruptedException e) {
System.out.println(name + " is interrupted.");
}
}
}
public static void main(String[] args) {
final Diner husband = new Diner(true, "husband");
final Diner wife = new Diner(true, "wife");
final Spoon sharedSpoon = new Spoon(wife);
Thread h = new Thread() {
#Override
public void run() {
husband.eatWith(wife, sharedSpoon);
}
};
h.start();
Thread w = new Thread() {
#Override
public void run() {
wife.eatWith(husband, sharedSpoon);
}
};
w.start();
try {
Thread.sleep(10000);
} catch (InterruptedException e) {
e.printStackTrace();
}
h.interrupt();
w.interrupt();
try {
h.join();
w.join();
} catch (InterruptedException e) {
e.printStackTrace();
}
}
}
package concurrently.deadlock;
import static java.lang.System.out;
/* This is an example of livelock */
public class Dinner {
public static void main(String[] args) {
Spoon spoon = new Spoon();
Dish dish = new Dish();
new Thread(new Husband(spoon, dish)).start();
new Thread(new Wife(spoon, dish)).start();
}
}
class Spoon {
boolean isLocked;
}
class Dish {
boolean isLocked;
}
class Husband implements Runnable {
Spoon spoon;
Dish dish;
Husband(Spoon spoon, Dish dish) {
this.spoon = spoon;
this.dish = dish;
}
#Override
public void run() {
while (true) {
synchronized (spoon) {
spoon.isLocked = true;
out.println("husband get spoon");
try { Thread.sleep(2000); } catch (InterruptedException e) {}
if (dish.isLocked == true) {
spoon.isLocked = false; // give away spoon
out.println("husband pass away spoon");
continue;
}
synchronized (dish) {
dish.isLocked = true;
out.println("Husband is eating!");
}
dish.isLocked = false;
}
spoon.isLocked = false;
}
}
}
class Wife implements Runnable {
Spoon spoon;
Dish dish;
Wife(Spoon spoon, Dish dish) {
this.spoon = spoon;
this.dish = dish;
}
#Override
public void run() {
while (true) {
synchronized (dish) {
dish.isLocked = true;
out.println("wife get dish");
try { Thread.sleep(2000); } catch (InterruptedException e) {}
if (spoon.isLocked == true) {
dish.isLocked = false; // give away dish
out.println("wife pass away dish");
continue;
}
synchronized (spoon) {
spoon.isLocked = true;
out.println("Wife is eating!");
}
spoon.isLocked = false;
}
dish.isLocked = false;
}
}
}
Example:
Thread 1
top:
lock(L1);
if (try_lock(L2) != 0) {
unlock(L1);
goto top;
Thread 2
top:
lock(L2);
if (try_lock(L1) != 0) {
unlock(L2);
goto top;
The only difference is Thread 1 and Thread 2 try to acquire the locks in a different order. Livelock could happen as follows:
Thread 1 runs acquires L1, then a context switch occurs. Thread 2 runs acquires L2, then another context switch occurs. Thread 1 runs and cannot acquire L2, but before releasing L1 a context switch occurs. Thread 2 runs and cannot acquire L1, releases L2, and a context switch occurs. Thread 1 releases L1, and now we are basically back to the starting state, and in theory these steps could keep repeating forever.