March 2009 - jboss-dev-forums - Jboss List Archives

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: PingStressTests hanging

by clebert.suconic＠jboss.com

Trustin, If I could have some help with this: I've been running org.jboss.messaging.tests.stress.remote.PingStressTest (change getNumberOfIterations to 100). One thread will fail with Ping, while other threads are also trying to close the connections. The Pinger thread will be waiting the Executors to shutdown, but a thread will stay alive forever. (Look at my previous post). Looking at NioWorker, I can see that selector.keys() is not empty hence the shutdown check is never made. I'm not sure if there is something that should be closed after the ping failure. If you want to verify what thread is holding the executor, this is the patch for the hack I've done here locally. (which I don't intend to commit it.. it's just for debug purposes): | Index: src/main/org/jboss/messaging/utils/JBMThreadFactory.java | =================================================================== | --- src/main/org/jboss/messaging/utils/JBMThreadFactory.java (revision 6014) | +++ src/main/org/jboss/messaging/utils/JBMThreadFactory.java (working copy) | @@ -37,7 +37,31 @@ | | private final AtomicInteger threadCount = new AtomicInteger(0); | | + private final ConcurrentSet<Thread> threads = new ConcurrentHashSet<Thread>(); | + | + private final String groupName; | + | + | + public void debug() | + { | + if (threads.isEmpty()) | + { | + System.out.println("Thread factory is empty"); | + } | + for (Thread t : threads) | + { | + System.out.println("This thread is still created by the factory : " + t + " name = " + t.getName()); | + StackTraceElement elements[] = t.getStackTrace(); | + for (StackTraceElement trace : elements) | + { | + System.out.println("Trace : " + trace); | + } | + } | + } | + | + | private final int threadPriority; | + | public JBMThreadFactory(final String groupName) | { | this(groupName, Thread.NORM_PRIORITY); | @@ -47,11 +71,27 @@ | { | group = new ThreadGroup(groupName + "-" + System.identityHashCode(this)); | threadPriority = priority; | + this.groupName = groupName; | } | | public Thread newThread(final Runnable command) | { | - Thread t = new Thread(group, command, "Thread-" + threadCount.getAndIncrement() + | + Runnable myRunnable = new Runnable() | + { | + public void run() | + { | + try | + { | + command.run(); | + } | + finally | + { | + threads.remove(Thread.currentThread()); | + } | + } | + }; | + | + Thread t = new Thread(group, myRunnable, "Thread-" + threadCount.getAndIncrement() + | " (group:" + | group.getName() + | ")"); | @@ -59,6 +99,7 @@ | // Don't want to prevent VM from exiting | t.setDaemon(true); | t.setPriority(threadPriority); | + threads.add(t); | return t; | } | } | Index: src/main/org/jboss/messaging/integration/transports/netty/NettyConnector.java | =================================================================== | --- src/main/org/jboss/messaging/integration/transports/netty/NettyConnector.java (revision 6014) | +++ src/main/org/jboss/messaging/integration/transports/netty/NettyConnector.java (working copy) | @@ -90,7 +90,10 @@ | public class NettyConnector implements Connector | { | // Constants ----------------------------------------------------- | + | + static JBMThreadFactory workerFactory = new JBMThreadFactory("jbm-netty-connector-worker-threads"); | | + | private static final Logger log = Logger.getLogger(NettyConnector.class); | | // Attributes ---------------------------------------------------- | @@ -236,7 +239,7 @@ | return; | } | | - workerExecutor = Executors.newCachedThreadPool(new JBMThreadFactory("jbm-netty-connector-worker-threads")); | + workerExecutor = Executors.newCachedThreadPool(workerFactory); | if (useServlet) | { | bossExecutor = Executors.newCachedThreadPool(new JBMThreadFactory("jbm-netty-connector-boss-threads")); | @@ -360,10 +363,14 @@ | { | try | { | - if (workerExecutor.awaitTermination(1, TimeUnit.SECONDS)) | + if (workerExecutor.awaitTermination(5, TimeUnit.SECONDS)) | { | break; | } | + else | + { | + workerFactory.debug(); | + } | } | catch (InterruptedException e) | { | View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215504#4215504 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215504

17 years, 1 month

1
0
0 / 0

[Design of POJO Server] - Re: JSR-299 servlet injection

by scott.stark＠jboss.org

jbossweb already calls out to the org.jboss.web.tomcat.service.TomcatInjectionContainer to process a servlet class, and this handles the current javaee5 annotations. That would need to be updated to recognize the jsr299 annotations, or just call out to a WebBeansManager if that is what should be done. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215501#4215501 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215501

17 years, 1 month

1
0
0 / 0

[Design of POJO Server] - Re: JSR-299 servlet injection

by pete.muir＠jboss.org

https://jira.jboss.org/jira/browse/WBINT-5 View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215495#4215495 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215495

17 years, 1 month

1
0
0 / 0

[Design of POJO Server] - JSR-299 servlet injection

by pete.muir＠jboss.org

Section 5.12.3 of the JSR-299 specification requires that JSR-299 DI occurs for Servlets. As there are no specified hooks in Servlet to perform a post construct lifecycle callback, I have exposed an API in Web Beans which allows the server to request a servlet is injected. This API is https://svn.jboss.org/repos/webbeans/ri/trunk/webbeans-ri-spi/src/main/ja... and is documented in the Web Beans documentation http://docs.jboss.org/webbeans/reference/snapshot/en-US/html/ri-spi.html#... We now need to get JBoss AS to perform servlet injection, and I would like guidance on the best way to hook this in. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215494#4215494 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215494

17 years, 1 month

1
0
0 / 0

[Design of JBoss Tools (dev)] - Re: Force full or partial update in vpe plugin

by kukeltje

cool... overriding getNodeForUptate works... thanks..... View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215493#4215493 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215493

17 years, 1 month

1
0
0 / 0

[Design of JBoss Tools (dev)] - Re: Force full or partial update in vpe plugin

by max.andersen＠jboss.com

hmm - report that in jira so we won't forget - thanks View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215492#4215492 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215492

17 years, 1 month

1
0
0 / 0

[Design of JBoss Tools (dev)] - Re: Force full or partial update in vpe plugin

by kukeltje

uuuhhhhhhhmmmmmmmmm guys, have you seen that the in the VpeAbstractTemplate the method is called getNodeForUptate (with a t instead of a d) that is why I implemented things wrongly I think. Refactoring would be wise ;-) Now lets check if that does work :-) View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215490#4215490 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215490

17 years, 1 month

1
0
0 / 0

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: PingStressTests hanging

by clebert.suconic＠jboss.com

I added some hacking to JBMFactory, to print threads that are still alive... And this thread is the one holding the Executor: This thread is still created by the factory : Thread[New I/O client worker #1268-1,5,jbm-netty-connector-worker-threads-29494967] name = New I/O client worker #1268-1 | Trace : org.jboss.netty.channel.socket.nio.NioWorker.run(NioWorker.java:169) | Trace : org.jboss.netty.util.ThreadRenamingRunnable.run(ThreadRenamingRunnable.java:72) | Trace : java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:885) | Trace : java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:907) | Trace : org.jboss.messaging.utils.JBMThreadFactory$1.run(JBMThreadFactory.java:88) | Trace : java.lang.Thread.run(Thread.java:619) | View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215489#4215489 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215489

17 years, 1 month

1
0
0 / 0

[Design of JBoss Tools (dev)] - Re: Force full or partial update in vpe plugin

by kukeltje

ok, I've been looking into this... When I select the parent node that needs updating myself and pass that into updateNode like | Node updateNode = ComponentUtil.getParentElementByName(sourceElement, XForms.TAG_REPEAT); | pageContext.getVisualBuilder().updateNode(updateNode); My whole page becomes empty. I do this right before I return from the create method in my template. Any other suggestions? View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215484#4215484 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215484

17 years, 1 month

1
0
0 / 0

[Design of POJO Server] - Re: Stable deployment API for tools

by alesj

Sure, I can try to port the changes back to 5_0. And just implement proper facade on top for existing 5.0 profile service integration. View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4215480#4215480 Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4215480

17 years, 1 month

1
0
0 / 0

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

jboss-dev-forums March 2009