Develop Reference

How can I prevent my UDP keepalive from using 100% CPU?

I have a Runnable that watches for data to send out UDP as well as to send a keep alive every 10 seconds. The process is taking 100% CPU. I tried setting the thread to low priority but didn't seem to make any difference.
private Runnable keepAliveRunnable = new Runnable() {
long nextSend = 0;
byte[] sendData;
#Override
public void run() {
if(DEBUG)
System.out.println("Starting keepAlive.");
while (socket != null) {
synchronized (socketLock) {
try {
sendData = sendQueue.poll();
if (sendData != null) {
socket.send(new DatagramPacket(sendData, sendData.length,
InetAddress.getByName(Main.ipAddress), 10024));
} else if (nextSend < System.currentTimeMillis()) {
if(DEBUG && nextSend < System.currentTimeMillis())
System.out.println("Update keepAlive.");
// Send /xremote
socket.send(new DatagramPacket(("/xremote").getBytes(),
("/xremote").getBytes().length,
InetAddress.getByName(Main.ipAddress), 10024));
nextSend = System.currentTimeMillis() + keepAliveTimeout;
// Send /info
socket.send(new DatagramPacket(("/info").getBytes(),
("/info").getBytes().length,
InetAddress.getByName(Main.ipAddress), 10024));
}
} catch (IOException e) {
e.printStackTrace();
if(!e.getMessage().contains("Socket closed")) {
e.printStackTrace();
}
}
}
}
System.out.println("keepAliveRunnable ended.");
}
};

Make sendQueue a LinkedBlockingQueue, and use poll timeouts.
You are busy waiting, which essentially forces your app to keep running the same logic over and over instead of giving the CPU back to the system.
Don't count on your own implementation of checking the time, that is unreliable and can result in what you're seeing. Instead, use blockingQueue.poll(10, TimeUnit.SECONDS), which automatically handles returning the CPU to the system.
I made a few other changes to your code; I put the duplicated packet construction code in a separate method, and I wrapped the synchronization of the socket only when the socket is actually being used. Notice how much cleaner it is when you let the queue do the work for you.
while(socket != null) {
try {
sendData = sendQueue.poll(10, TimeUnit.SECONDS);
if (sendData != null) {
sendPacket(sendData);
} else {
sendPacket("/xremote".getBytes());
sendPacket("/info".getBytes());
}
} catch (IOException e) {
e.printStackTrace();
if (!e.getMessage().contains("Socket closed")) {
e.printStackTrace();
}
}
}
And here's sendPacket:
private static void sendPacket(byte[] data) throws UnknownHostException, IOException {
// Note, you probably only have to do this once, rather than looking it up every time.
InetAddress address = InetAddress.getByName(Main.ipAddress);
DatagramPacket p = new DatagramPacket(data, data.length, address, 10024);
synchronized(socketLock) {
socket.send(p);
}
}

You should add a Thread.sleep() at the bottom of your while loop, to slow down your loop. As is, you're busy-waiting and churning the CPU while you wait for the nextSend time to be reached. Thread.sleep() will actually pause the thread, allowing other threads and processes to use the CPU while this one sleeps.
Sleeping for a 10th of a second (100 milliseconds) should be a good amount of time to sleep between iterations of your loop, if your goal is to actually do work every 10 seconds.
There are more advanced techniques for dispatching work every so often, like ScheduledExecutorService, which you could also consider using. But for a small application the pattern you're using is fine, just avoid busy waiting.

I think rather than polling your sendqueue, its better to use semaphore signal and wait.
When a packet is inserted in sendqueue, call semaphore signal.
Use semaphore wait instead of call to sendqueue.poll().
I assume you have separate threads for pushing popping data from sendqueue.
This is standard consumer producer problem. https://en.wikipedia.org/wiki/Producer%E2%80%93consumer_problem

After digging through my code, I had realized that over time I had whittled down the number of processes sending data to 1 (duh) so I really didn't need the runnable as I could just send the data directly. I also set up a separate runnable and used ScheduledExecutor. I thought I would just put that here for other to see. Durron597's code is a little prettier but since I'm only sending two packs now I decided to just put the code together.
// In main
pingXAir();
private void pingXAir() {
System.out.println("Start keepAlive");
ScheduledExecutorService executorService = Executors.newScheduledThreadPool(1);
executorService.scheduleAtFixedRate(keepAliveRunnable, 0, 5, TimeUnit.SECONDS);
}
private Runnable keepAliveRunnable = new Runnable() {
#Override
public void run() {
synchronized (socketLock) {
try {
if (DEBUG)
System.out.println("Update keepAlive.");
// Send /xremote
socket.send(new DatagramPacket(("/xremote").getBytes(),
("/xremote").getBytes().length,
InetAddress.getByName(Main.ipAddress), 10024));
// Send /info
socket.send(new DatagramPacket(("/info").getBytes(),
("/info").getBytes().length,
InetAddress.getByName(Main.ipAddress), 10024));
} catch (IOException e) {
e.printStackTrace();
if (!e.getMessage().contains("Socket closed")) {
e.printStackTrace();
}
}
}
}
};

Where to use Thread interupt

I have some old code I am working with, and I'm not too experienced with Threads (mostly work on the front end). Anyway, this Thread.sleep is causing the thread to hang and I'm unsure what to do about it. I thought about using a counter and throwing a Thread.currentThread.interupt, but unsure of where to put it or which thread it will interupt. Here is an example of the dump. As you can see the thread count is getting pretty high at 1708.
Any advice?
"Thread-1708" prio=6 tid=0x2ceec400 nid=0x2018 waiting on condition
[0x36cdf000] java.lang.Thread.State: TIMED_WAITING (sleeping) at
java.lang.Thread.sleep(Native Method) Locked ownable synchronizers:
- None "Thread-1707" prio=6 tid=0x2d16b800 nid=0x215c waiting on condition [0x36c8f000] java.lang.Thread.State: TIMED_WAITING
(sleeping) at java.lang.Thread.sleep(Native Method) Locked ownable
synchronizers:
- None
#Override
public void run()
{
Connection con = null;
int i = 0;
while (is_running)
{
try
{
con = ConnectionManager.getConnection();
while (!stack.isEmpty())
{
COUNT++;
String line = (String) stack.pop();
getPartMfr(line);
try
{
if (this.mfr != null && !this.mfr.equals(EMPTY_STR))
{
lookupPart(con, line);
}
}
catch (SQLException e)
{
e.printStackTrace();
}
if (COUNT % 1000 == 0)
{
Log log = LogFactory.getLog(this.getClass());
log.info("Processing Count: " + COUNT);
}
}
}
catch (NamingException e)
{
e.printStackTrace();
}
catch (SQLException e)
{
e.printStackTrace();
}
finally
{
try
{
ConnectionManager.close(con);
}
catch (SQLException e)
{
e.printStackTrace();
}
}
try {
Thread.sleep(80);
} catch (InterruptedException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
this.finished = true;
}
Here is where it calls the run method, as you can see it does set it to false, but I guess it is missing threads?
HarrisWorker w[] = new HarrisWorker[WORKER_POOL_SIZE];
try
{
for (int i = 0; i < w.length; i++)
{
w[i] = new HarrisWorker(pw);
w[i].start();
}
pw.println(headers());
File inputDir = new File(HARRIS_BASE);
String files[] = inputDir.list();
for (String file : files)
{
try
{
File f = new File(HARRIS_BASE + File.separator + file);
if (f.isDirectory())
continue;
final String workFile = workDir + File.separator + file;
f.renameTo(new File(workFile));
FileReader fr = new FileReader(workFile);
BufferedReader br = new BufferedReader(fr);
String line = br.readLine();
boolean firstLine = true;
while (line != null)
{
if (firstLine)
{
firstLine = false;
line = br.readLine();
continue;
}
if (line.startsWith(","))
{
line = br.readLine();
continue;
}
// if(line.indexOf("103327-1") == -1)
// {
// line = br.readLine();
// continue;
// }
HarrisWorker.stack.push(line);
line = br.readLine();
}
br.close();
fr.close();
for (int i = 0; i < w.length; i++)
{
w[i].is_running = false;
while (!w[i].finished)
{
Thread.sleep(80);
}
}
move2Processed(file, workFile);
long etime = System.currentTimeMillis();
System.out.println("UNIQUE PARTS TOTAL FOUND: " + HarrisWorker.getFoundCount() + " of " + HarrisWorker.getUniqueCount() + ", "
+ (HarrisWorker.getFoundCount() / HarrisWorker.getUniqueCount()));
System.out.println("Time: " + (etime - time));
}
catch (Exception e)
{
e.printStackTrace();
File f = new File(workDir + File.separator + file);
if (f.exists())
{
f.renameTo(new File(HARRIS_BASE + File.separator + ERROR + File.separator + file));
}
}
}
}

As a direct answer to the question in your title - nowhere. There is nowhere in this code that needs a Thread.interrupt().
The fact that the thread name is Thread-1708 does not necessarily mean there are 1708 threads. One can choose arbitrary names for threads. I usually include the name of the executor or service in the thread name. Maybe 1600 are now long stopped and there are only around a hundred alive. Maybe this particular class starts naming at 1700 to distinguish from other uses.
1708 threads may not be a problem. If you have a multi-threaded server that is serving 2000 connections in parallel, then it certainly expectable that there are 2000 threads doing that, along with a bunch of other threads.
You have to understand why the sleep is there and what purpose it serves. It's not there to just hog memory for nothing.
Translating the code to "plaintext" (btw it can be greatly simplified by using try-with-resources to acquire and close the connection):
Acquire a connection
Use the connection to send (I guess) whatever is in the stack
When failed or finished - wait 80ms (THIS is your sleep)
If run flag is still set - repeat from step 1
Finish the thread.
Now reading through this, it's obvious that it's not the sleep that's the problem. It's that the run flag is not set to false. And your thread just continues looping, even if it can't get the connection at all - it will simply spend most of its time waiting for the retry. In fact - even if you completely strip the sleep out (instead of interrupting it mid-way), all you will achieve is that the Threads will start using up more resources. Given that you have both a logger and you print to stdout via printStackTrace, I would say that you have 2 problems:
Something is spawning threads and not stopping them afterwards (not setting their run flag to false when done)
You are likely getting exceptions when getting the Connection, but you never see them in the log.
It might be that the Thread is supposed to set it's own run flag (say when the stack is drained), but you would have to decide that yourself - that depends on a lot of specifics.

Not an answer but some things you should know if you are writing code for a live, production systemn:
:-( Variable and method both have the same name, run. A better name for the variable might be keep_running Or, change the sense of it so that you can write while (! time_to_shut_down) { ... }
:-( Thread.sleep(80) What is this for? It looks like a big red flag to me. You can never fix a concurrency bug by adding a sleep() call to your code. All you can do is make the bug less likely to happen in testing. That means, when the bug finally does bite, it will bite you in the production system.
:-( Your run() method is way too complicated (the keyword try appears four times). Break it up, please.
:-( Ignoring five different exceptions catch (MumbleFoobarException e) { e.printStackTrace(); } Most of those exceptions (but maybe not the InterruptedException) mean that something is wrong. Your program should do something more than just write a message to the standard output.
:-( Writing error messages to standard output. You should be calling log.error(...) so that your application can be configured to send the messages to someplace where somebody might actually see them.

java code to wait for parallel code to finish

I m having a server code to process an image.
Now there are n number of requests which tries to execute the code which results in OutOfMemory error or the server to hang and the server goes to not responding state.
To stop the code from executing at once all the requests I m limiting to execute the code one at a time using the below method where i have a variable
if the variable is 10 then wait for the variable to come at 0
if at 0 then set it to 10 then execute the code
run the code and finally set i to 0
The code here -
static newa.Counter cn;
public int getCounta() {
return cn.getCount();
}
public void setCounta(int i) {
cn = new newa.Counter();
cn.setCount(i);
}
at the function i m doing this -
public BufferedImage getScaledImage(byte[] imageBytes)
{
int i=0;
Boolean b = false;
BufferedImage scaledImage = null;
newa.NewClass1 sc = new newa.NewClass1();
try {
sc.getCounta();
} catch (NullPointerException ne) {
sc.setCounta(0);
}
i = sc.getCounta();
if(i==0)
{
sc.setCounta(10);
b = true;
}
else
{
while( b == false)
{
try
{
Thread.sleep(2000);
i = sc.getCounta();
if( i==0)
{
sc.setCounta(10);
b = true;
System.out.println("Out of Loop");
}
} catch (InterruptedException ex) {
System.out.println("getScaledImage Thread exception: " + ex);
}
}
}
..... execute further code
try { } catch { } finally { sc.setCounta(0); }
}
Is there any way I can have this simplified using the Runnable interface or something as I don't know how to do multi-threading.

Forget about the counter and use a synchronized method. Changed your method head to this:
public synchronized BufferedImage getScaledImage(byte[] imageBytes)
This lets all the threads entering the method wait until no other thread is executing the method.

If you want only a small number of threads doing the processing you can use Executor framework to have a thread pool of 10 threads. This will ensure that at one time maximum of 10 threads will be processing the requests.

ZeroMQ - jzmq .recvZeroCopy() fails to get any message while .recv() works

So I am writing my own piece of stuff using jzmq GIT master branch and ZeroMQ 3.2.3.
After installation I tried to test the following simple PUB/SUB program, where a publisher and a subscriber talk in a single process. Since the test is under Windows, I used TCP.
public class ZMQReadynessTest {
private ZMQ.Context context;
#Before
public void setUp() {
context = ZMQ.context(1);
}
#Test
public void testSimpleMessage() {
String topic = "tcp://127.0.0.1:31216";
final AtomicInteger counter = new AtomicInteger();
// _____________________________________ create a simple subscriber
final ZMQ.Socket subscribeSocket = context.socket(ZMQ.SUB);
subscribeSocket.connect(topic);
subscribeSocket.subscribe("TestTopic".getBytes());
Thread subThread = new Thread() {
#Override
public void run() {
while (true) {
String value = null;
// This would result in trouble /\/\/\/\/\/\/\/\/\
{
ByteBuffer buffer = ByteBuffer.allocateDirect(100);
if (subscribeSocket.recvZeroCopy( buffer,
buffer.remaining(),
ZMQ.DONTWAIT
) > 0 ) {
buffer.flip();
value = buffer.asCharBuffer().toString();
System.out.println(buffer.asCharBuffer().toString());
}
}
// This works perfectly + + + + + + + + + + + + +
/*
{
byte[] bytes = subscribeSocket.recv(ZMQ.DONTWAIT);
if (bytes == null || bytes.length == 0) {
continue;
}
value = new String(bytes);
}
*/
if (value != null && value.length() > 0) {
counter.incrementAndGet();
System.out.println(value);
break;
}
}
}
};
subThread.start();
// _____________________________ create a simple publisher
ZMQ.Socket publishSocket = context.socket(ZMQ.PUB);
publishSocket.bind("tcp://*:31216");
try {
Thread.sleep(3000); // + wait 3 sec to make sure its ready
} catch (InterruptedException e) {
e.printStackTrace();
fail();
}
// publish a sample message
try {
publishSocket.send("TestTopic".getBytes(), ZMQ.SNDMORE);
publishSocket.send("This is test string".getBytes(), 0);
subThread.join(100);
} catch (InterruptedException e) {
e.printStackTrace();
fail();
}
assertTrue(counter.get() > 0);
System.out.println(counter.get());
}
}
Now as you can see, in the subscriber if I use a simple .recv(ZMQ.DONTWAIT) method, it works perfectly. However, if I am using the direct byte buffer I got nothing returned - and I got the following exception, seems like on program exit:
Exception in thread "Thread-0" org.zeromq.ZMQException: Resource temporarily unavailable(0xb)
at org.zeromq.ZMQ$Socket.recvZeroCopy(Native Method)
at ZMQReadynessTest$1.run(ZMQReadynessTest.java:48)
I also tried to use a simple ByteBuffer (not a direct buffer), which doesn't throw the exception above; but also return me nothing.
Does anybody know how to resolve the above?
I don't want to create byte[] objects all around, as I am doing some high performance system. If this cannot be resolved, I might simply use Unsafe instead. But I really want to work in the "supposed way".
Thanks in advance.
Alex

Locking a process in Java

I have a Java servlet which calls another software (say S) over a TCP connection. This software S uses a network resource, and the output has to be retrived from a hyperlink(using wget).
Since it's the same hyperlink I need to download my result from (irrespective of the request), it results into incorrect results few requests. I basically need to lock the use of this network resource across different processes (I believe each call from the servlet is going to create a new process).
I tried to use ReentrantLock (but I guess it only works with threads and not accross processes).
Please let me know how can this be achieved.
Thanks

Here is how to do cross-process locking in Java. Adjust to your needs and add error/exception checking/handling as necessary.
// Tester
try {
if (crossProcessLockAcquire(SomeClassInYourApp.class, 3000)) {
// Success - This process now has the lock. (Don't keep it too long.)
}
else {
// Fail (Timeout) - Another process still had the lock after 3 seconds.
}
} finally {
crossProcessLockRelease(); // try/finally is very important.
}
// Acquire - Returns success ( true/false )
private static boolean crossProcessLockAcquire(final Class<?> c, final long waitMS) {
if (fileLock == null && c != null && waitMS > 0) {
try {
long dropDeadTime = System.currentTimeMillis() + waitMS;
File file = new File(lockTempDir, c.getName() + ".lock");
RandomAccessFile randomAccessFile = new RandomAccessFile(file, "rw");
FileChannel fileChannel = randomAccessFile.getChannel();
while (System.currentTimeMillis() < dropDeadTime) {
fileLock = fileChannel.tryLock();
if (fileLock != null) {
break;
}
Thread.sleep(250); // 4 attempts/sec
}
} catch (Exception e) {
e.printStackTrace();
}
}
return fileLock == null ? false : true;
}
// Release
private static void crossProcessLockRelease() {
if (fileLock != null) {
try {
fileLock.release();
fileLock = null;
} catch (IOException e) {
e.printStackTrace();
}
}
}
// Some class vars and a failsafe lock release.
private static File lockTempDir = new File(System.getProperty("java.io.tmpdir") + File.separator + "locks");
private static FileLock fileLock = null;
static {
Runtime.getRuntime().addShutdownHook(new Thread() {
public void run(){
crossProcessLockRelease();
}
});
}

Why are you reusing this TCP connection? If it's easy to set up, just set one up every time you need it. For example, with an HTTP request, you should just make a new request every time.
My guess is that you have something static that shouldn't be, so multiple threads are using it when they should all have their own version.
If they're expensive, consider creating one-per-thread with ThreadLocal.
If even that doesn't work, and you don't mind threads blocking, just add "synchronized" to the method that's causing the problem.

The resource you are trying to lock has to support looking. It would be better if the service didn't need to be locked externally.
As a work around you can use a ServerSocket to lock a resource between processes.