Hi all,

Someone test it? it's only happening to me?


Regards!

2011/1/17 Oscar Pernas <os...@pernas.es>

> Hi Tim,
>
> Here you have the test case, is a visual studio 2005 project, you only have
> to compile giving an ip address in where there are no brokers. In this
> example when the connection with broker wont be stablished, the exception is
> catched and it calls to cleanup method to freeing resources. You are going
> to see that when you delete connection and session, the cpu start to take
> 100%.
>
>
> Tell me if you need more information.
>
>
>
> Oscar Pernas Plaza
>
> 2011/1/17 Timothy Bish <tabish...@gmail.com>
>
> On Mon, 2011-01-17 at 14:05 +0100, Oscar Pernas wrote:
>> > Hi all,
>> >
>> >
>> > I've found the problem. The problem was that when you make the
>> > connection->start(); if it is impossible to connect to broker,
>> ActiveMQ-cpp
>> > throws an exception. I catch this exception and call to cleanup method
>> to
>> > free all resources.
>> > In the moment that I'm deleting resources, particulary connection
>> pointer,
>> > the library call to the mutex.wait method with 15 seconds time and takes
>> the
>> > 100% usage of the CPU. Particulary in this disconnect:
>> >
>> >         // Now inform the Broker we are shutting down.
>> >         this->disconnect( lastDeliveredSequenceId );
>> >
>> >
>>
>> Without some code samples to show how you are using the library its hard
>> to make a determination about what is going on.  Can you provide a
>> sample app that demonstrates the issue?
>>
>> Regards
>>
>>
>>
>> > I saw that is the WaitForSIngleObject the method that produces the high
>> cpu
>> > load.
>> >
>> >
>> > If the connection cannot start, I have to make a close, but it hangs my
>> cpu,
>> > If I catch the exception could I set  connection to NULL or will be a
>> leak?
>> >
>> > Any other idea to close connection resource withouth this cpu high load?
>> >
>> > 2011/1/17 Oscar Pernas <os...@pernas.es>
>> >
>> > > Hi all,
>> > >
>> > >
>> > > I am developing an application that uses JMS and needs to start with
>> or
>> > > without the connection to broker was stablished or not. For this, I am
>> > > coding a reconnect thread that is spawned when the initial reconnects
>> fail.
>> > > I've set this connection strings:
>> > >
>> > > failover://(tcp://10.0.2.22:61616
>> > >
>> )?startupMaxReconnectAttempts=1&initialReconnectDelay=1&maxReconnectDelay=1
>> > >
>> > >
>> > > When the application starts and try to connect, if it fails, the
>> thread is
>> > > spawned and trying to reconnect every 30 seconds, with this case I can
>> see
>> > > two weirds behaviours.
>> > >
>> > > - Every time that the thread wants to reconnect, the process eats the
>> 100%
>> > > of the cpu. I've follow the code and I dont know why, but in the
>> function
>> > >
>> > > bool CountDownLatch::await( long long timeOut )
>> > >
>> > > I have a mutex wait timeOut that is equal to 15000 seconds.
>> > >
>> > >             mutex.wait( timeOut );
>> > >
>> > > Why the code is waiting 15000 seconds? why it takes the 100% of the
>> cpu?
>> > >
>> > > The await function is called by:
>> > >
>> > > Pointer<Response> ResponseCorrelator::request  -->  response =
>> > > futureResponse->getResponse( timeout );
>> > >
>> > >
>> > > Im developing in windows 2000.
>> > >
>> > >
>> > >
>> > > Regards
>> > >
>> > >
>> > >
>> > >
>> > > --
>> > > Óscar Pernas Plaza.
>> > >
>> >
>> >
>> >
>>
>> --
>> Tim Bish
>> ------------
>> FuseSource
>> Email: tim.b...@fusesource.com
>> Web: http://fusesource.com
>> Twitter: tabish121
>> Blog: http://timbish.blogspot.com/
>>
>>
>>
>
>
> --
> Óscar Pernas Plaza.
>



-- 
Óscar Pernas Plaza.

Reply via email to