[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Alexey Kazakov reassigned JBIDE-3461: ------------------------------------- Assignee: Rob Stryker -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Max Rydahl Andersen updated JBIDE-3461: --------------------------------------- Fix Version/s: 3.0.0.CR2 -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Toma�¾ Zupan commented on JBIDE-3461: --------------------------------------- I looked threads by new "Java VisualVM" tool and I can see that every time when I press the "Stop" button, some new threads are created for terminating JBOSS server and they are all terminated after a while when JBOSS server is stoped, except one which is fully active all the time. Here is a dump of that thread: 2009-01-14 08:03:30 Full thread dump Java HotSpot(TM) 64-Bit Server VM (11.0-b16 mixed mode): "Thread-50" prio=6 tid=0x000000002978f400 nid=0xd64 runnable [0x000000003342f000..0x000000003342fa10] java.lang.Thread.State: RUNNABLE at org.jboss.ide.eclipse.as.core.server.internal.launch.StopLaunchConfiguration.stop(StopLaunchConfiguration.java:60) at org.jboss.ide.eclipse.as.core.server.internal.JBossServerBehavior$1.run(JBossServerBehavior.java:84) Locked ownable synchronizers: - None ... And every time when I stop the server this kind of thread is created (name is "Thread-" + number) and never terminate, so the machine is slower and slower. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Rob Stryker commented on JBIDE-3461: ------------------------------------ Looking at the code, this seems very weird. The code is as follows: ILaunchConfigurationWorkingCopy wc = createLaunchConfiguration(server); ILaunch launch = wc.launch(ILaunchManager.RUN_MODE, new NullProgressMonitor()); IProcess stopProcess = launch.getProcesses()[0]; while( !stopProcess.isTerminated()) {} return stopProcess.getExitValue() == 0 ? true : false; The only way this block of code is staying open is if the Stop process never ends. Since the stop process is launched via a launch configuration, the stop process is actually an external process we just have a reference to. It's not run within the eclipse process. So... you may want to check your system to see if a bunch of rogue java processes are still going. If you're seeing 20 or more threads stuck on that loop, do you also see 20 or more java processes running on your system? The stop process, specifically, is basically a call to ${server home}/bin/shutdown.jar. This is almost always a very very short block of code. In fact, looking at THAT class, there is only 1 loop in the class at all, and that is to parse the arguments! I find it hard to believe this is failing. And from the code above, once the stop process terminates, the loop should end. I'm committing a change to make sure the loop doesn't keep going even after the server is stopped. But I'm not sure it'll fix your problem entirely. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Toma�¾ Zupan commented on JBIDE-3461: --------------------------------------- I'm looking that process in task manager and I can see that that "javaw.exe" process is created on server start and disappear after server is stoped. So process is successfully terminated but those thread is still active. I can imagine that something is wrong with isTerminated() method or the stopProcess is not the right one? -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Toma�¾ Zupan commented on JBIDE-3461: --------------------------------------- I was so curious what is wrong here that I couldn't keep back. So I decided to learn how to develope JBossTools. After installing, studying and debuging I find what is wrong. The problem is in threading, more precisely in Memory Consistency Errors (look http://java.sun.com/docs/books/tutorial/essential/concurrency/memconsist....). Because writing and reading of the fTerminated variable in RuntimeProcess class is from the two different threads, memory consistency error is happend. So the method stopProcess.isTerminated() is returnig false all the time altough ProcessMonitor thread is already set the fTerminated variable to true. The solution is to synchronize the reading and writing of the fTerminated variable which is actually an eclipse bug. Here in my case it is enough to put some thread function (like Thread.yield() or Thread.sleep(0)) in while loop and stopProcess.isTerminated() would return true. But by Java specification this is not guaranteed and could emerged in some other JVM. The other problem here is that empty while loop wastes processor times. On link http://java.sun.com/docs/books/tutorial/essential/concurrency/guardmeth.html you can see some suggestions how to solve this problem. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Rob Stryker reopened JBIDE-3461: -------------------------------- thanks for the input, tomaz. I'll make the changes when I can. Since I already added another loop-control, this is no longer critical, but i'l re-opening so I don't forget to make the changes -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Rob Stryker commented on JBIDE-3461: ------------------------------------ Actually, maybe I can get rid of this entirely by instead executing just a JMX command. The shutdown script (shutdown.bat, shutdown.jar) already assumes that JMX is up and can receive the shutdown call. If I simply execute a JMX shutdown command, I can stop from spawning another process at all, and then also not have to do any loop whatsoever. I can look into changing the shutdown into this type of method. I think it'd make things much more efficient. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Max Rydahl Andersen updated JBIDE-3461: --------------------------------------- Fix Version/s: 3.0.0.CR2 (was: 3.0.0.GA) this sounds like something we don't want to wait for GA to fix (please correct me if am wrong). Switching to JMX sounds like a good thing, but do we really want to remove the option for using simple start and stop process ? (Jmx requires host, port etc.) -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Max Rydahl Andersen commented on JBIDE-3461: -------------------------------------------- sorry, thought it was just a "simple" kill process. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Max Rydahl Andersen commented on JBIDE-3461: -------------------------------------------- Great and big thanks to Tomaz to spot the problem - much appreciated! -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBIDE-3461?page=com.atlassian.jira.plu... ] Max Rydahl Andersen commented on JBIDE-3461: -------------------------------------------- hehe - feel free to hunt down more ;) -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira