Mailing-List: contact dev-help@activemq.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@activemq.apache.org
Date: Mon, 4 Feb 2013 17:22:12 +0000 (UTC)
From: "Helen Huang (JIRA)" <jira@apache.org>
To: dev@activemq.apache.org
Message-ID: <JIRA.12630466.1359760568868.240337.1359998532902@arcas>
In-Reply-To: <JIRA.12630466.1359760568868@arcas>
References: <JIRA.12630466.1359760568868@arcas>
Subject: [jira] [Commented] (AMQCPP-451) The number of CMS threads increases
 when we restart the message broker repeatedly
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/AMQCPP-451?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13570406#comment-13570406 ] 

Helen Huang commented on AMQCPP-451:
------------------------------------

Have you tried to use multiple senders and receivers in your test? I have 10 pairs of senders and receivers in my test. In each sender-receiver pair, the sender sends to a topic that the receiver receives from. All sender and receiver open their own threads and use CMSTemplate to send and receive messages.

A sender thread looks like:
for(...)
{
  sender->sendMessage(...);//URL is "tcp://localhost:61616?connection.sendTimeout=1000"
  Sleep(50);
}

A receiver thread looks like:
for (...)
{
  receiver->receiveMessage(...); //URL is "tcp://localhost:61616". Receive timeout is 200
}

Also I did the test through a com interface. I will test again without it and see if I can reproduce.

                
> The number of CMS threads increases when we restart the message broker repeatedly 
> ----------------------------------------------------------------------------------
>
>                 Key: AMQCPP-451
>                 URL: https://issues.apache.org/jira/browse/AMQCPP-451
>             Project: ActiveMQ C++ Client
>          Issue Type: Bug
>    Affects Versions: 3.5.0
>         Environment: Windows xp service pack 3, ActiveMQ broker 5.3.1, apr 1.4.2, apr-util 1.3.9, apr iconv 1.2.1
>            Reporter: Helen Huang
>            Assignee: Timothy Bish
>            Priority: Critical
>             Fix For: 3.5.0
>
>         Attachments: CrashHang_Report__CMHMultiTestWithCom.exe__02012013222736766.mht
>
>
> The number of CMS threads increases when we restart the message broker repeatedly. We have ten senders and ten receivers in out test. Every sender sends out 10-20 messages every second. We restart the broker every 1 minute. The number of threads in CMS grew from about 50 to 600 after about 15 hours of testing. The attached dump was created after we ran two hours of test. From the dump we find a large number of threads are stuck at the following place:
> Function   Source 
> ntdll!KiFastSystemCallRet    
> ntdll!ZwWaitForSingleObject+c    
> kernel32!WaitForSingleObjectEx+a8    
> kernel32!WaitForSingleObject+12    
> activemq_cppud!decaf::internal::util::concurrent::PlatformThread::interruptibleWaitOnCondition+25    
> activemq_cppud!`anonymous namespace'::threadEntryMethod+9b    
> msvcr80d!_callthreadstartex+51    
> msvcr80d!_threadstartex+87    
> kernel32!BaseThreadStart+37 

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira