AppletTalk.com Java discussions newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  SUN 1.4.x JVM: Windows Event handle leak          AppletTalk.com Forum Index -> JVM, native methods and hardware View previous topic :: View next topic   Author Message Stephen Berard Guest Posted: Tue Dec 09, 2003 5:10 pm    Post subject: SUN 1.4.x JVM: Windows Event handle leak I have an application that is leaking handles when running under the Sun 1.4.x JVMs on Windows XP Pro. Running under 1.3.1 the same code does not leak. Under further investigation it appears that the handles are Windows Event handles. This can be seen in the Windows Task Manager and Performance Manager. I have narrowed it down to a section of code that performs a polling loop (code is too lengthy to post). The polling loop is a synchronized queue. Items are popped off the queue and processed. The results are returned in a result object (whose access methods are synchronized). Then the item is placed back on the end of the queue. There is a pause (2-3 seconds) between the processing of each item. The processing of each item involves a JNI call to obtained data via USB. (This code has not been modified and runs correctly under 1.3.x JVMs). The strange thing is that handles only leak occasionally. If I leave my app running for a few hours I may leak 5-10 handles. This has no correlation to the number of items processed or the number of times through the full poll loop. The only clue I have is that the garbage collector seems to be running when a handle leak occurs. However, a handle leak does not occur on every GC cycle. I suspect that this problem has something to do with synchronized objects. What's weird is that this same exact code runs perfectly on 1.3.1. Has anyone else seen similar behavior? Any clues on how to debug this? Thanks for your help, Stephen Back to top Michael Amling Guest Posted: Thu Dec 11, 2003 2:49 pm    Post subject: Re: SUN 1.4.x JVM: Windows Event handle leak Stephen Berard wrote: Quote: I have an application that is leaking handles when running under the Sun 1.4.x JVMs on Windows XP Pro. Running under 1.3.1 the same code does not leak. Under further investigation it appears that the handles are Windows Event handles. This can be seen in the Windows Task Manager and Performance Manager. I have narrowed it down to a section of code that performs a polling loop (code is too lengthy to post). The polling loop is a synchronized queue. Items are popped off the queue and processed. The results are returned in a result object (whose access methods are synchronized). Then the item is placed back on the end of the queue. There is a pause (2-3 seconds) between the processing of each item. The processing of each item involves a JNI call to obtained data via USB. (This code has not been modified and runs correctly under 1.3.x JVMs). The strange thing is that handles only leak occasionally. If I leave my app running for a few hours I may leak 5-10 handles. This has no correlation to the number of items processed or the number of times through the full poll loop. The only clue I have is that the garbage collector seems to be running when a handle leak occurs. However, a handle leak does not occur on every GC cycle. I suspect that this problem has something to do with synchronized objects. What's weird is that this same exact code runs perfectly on 1.3.1. Has anyone else seen similar behavior? Any clues on how to debug this? Can you define the term "leak"? What symptoms lead you to the conclusion that "handles" are "leaking"? --Mike Amling Back to top Stephen Berard Guest Posted: Thu Dec 11, 2003 7:37 pm    Post subject: Re: SUN 1.4.x JVM: Windows Event handle leak Michael Amling <nospam (AT) nospam (DOT) com> wrote Quote: Stephen Berard wrote: I have an application that is leaking handles when running under the Sun 1.4.x JVMs on Windows XP Pro. Running under 1.3.1 the same code does not leak. Under further investigation it appears that the handles are Windows Event handles. This can be seen in the Windows Task Manager and Performance Manager. I have narrowed it down to a section of code that performs a polling loop (code is too lengthy to post). The polling loop is a synchronized queue. Items are popped off the queue and processed. The results are returned in a result object (whose access methods are synchronized). Then the item is placed back on the end of the queue. There is a pause (2-3 seconds) between the processing of each item. The processing of each item involves a JNI call to obtained data via USB. (This code has not been modified and runs correctly under 1.3.x JVMs). The strange thing is that handles only leak occasionally. If I leave my app running for a few hours I may leak 5-10 handles. This has no correlation to the number of items processed or the number of times through the full poll loop. The only clue I have is that the garbage collector seems to be running when a handle leak occurs. However, a handle leak does not occur on every GC cycle. I suspect that this problem has something to do with synchronized objects. What's weird is that this same exact code runs perfectly on 1.3.1. Has anyone else seen similar behavior? Any clues on how to debug this? Can you define the term "leak"? What symptoms lead you to the conclusion that "handles" are "leaking"? --Mike Amling Using PerfMon or the Task Manager the number of handles increases over time. This does not happen with the 1.3x JVMs, they remain relatively flat. I did a bit of investigation and what's happening is that the JVM is creating a number of unname NT Events. I have see a few posting on Sun's Java forum from people having similar issues. Here are some links: http://forum.java.sun.com/thread.jsp?forum=37&thread=439546 http://forum.java.sun.com/thread.jsp?forum=4&thread=452929 Just to be clear, I'm not 100% sure that this is a leak. The handle count goes up by a random amount; it does not seem to be related to anything in my program. I have tried polling more, polling less, adding items to the poll list, make the poll list smaller, etc. It looks to me like a race condition. Any help is appreciated. Cheers, Stephen Back to top Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First         AppletTalk.com Forum Index -> JVM, native methods and hardware All times are GMT Page 1 of 1   Jump to: Select a forum Java Newsgroups----------------Java DiscussionsWeb-based java games comp.lang.java 3D Graphics API's for Java Java: Support and criticism Java System Announcements comp.lang.java.api JavaBeans Java and CORBA Java and Databases comp.lang.java.developer Java GUI Toolkits Java Help comp.lang.java.javascript JVM, native methods and hardware comp.lang.java.misc Java Language Programming Security and Java comp.lang.java.setup Java Software Tools comp.lang.java.tech ibm.software.java.linux  Non-English Java Newsgroups----------------Java Language (German)Language Java (Français)Java (Italian)  You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum Powered by phpBB © 2001, 2006 phpBB Group SEO toolkit © 2004-2006 webmedic.