jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From João Tiago Ferreira <joao.tiago.ferre...@novabase.pt>
Subject RE: JMS Assertion??
Date Tue, 22 Jun 2010 15:35:05 GMT


> -----Original Message-----
> From: sebb [mailto:sebbaz@gmail.com]
> Sent: terça-feira, 22 de Junho de 2010 16:22
> To: JMeter Users List
> Subject: Re: JMS Assertion??
> 
> On 22/06/2010, João Tiago Ferreira <joao.tiago.ferreira@novabase.pt>
> wrote:
> >
> >
> >  > -----Original Message-----
> >  > From: sebb [mailto:sebbaz@gmail.com]
> >  > Sent: terça-feira, 22 de Junho de 2010 11:57
> >  > To: JMeter Users List
> >  > Subject: Re: JMS Assertion??
> >  >
> >
> > > On 21/06/2010, João Tiago Ferreira
> <joao.tiago.ferreira@novabase.pt>
> >  > wrote:
> >  > > Hi
> >  > >
> >  > >  Patch for JMS Queue receiver in
> >  > https://issues.apache.org/bugzilla/show_bug.cgi?id=49482
> >  >
> >  > Thanks, I've had a brief look at the code.
> >  >
> >  > >  The great difference from Topic Subscriber is the cleaning of
> >  > resources and is impossibility of choosing the client method to
> use.
> >  >
> >  > What do you mean - choosing the client method to use?
> >
> >
> >
> http://jakarta.apache.org/jmeter/usermanual/component_reference.html#JM
> S_Subscriber
> >
> >  Choose between Receiver.receive() and MessageListener.onMessage(). I
> use receive()
> >
> 
> OK, understood.
> 
> I'm not sure why the Topic Subscriber allows the choice anyway - who
> cares how the message is received?
> 

Dont't know...

> >  >
> >  > >  Instead of cleaning the Receivers in the end of the test they
> are
> >  > cleaned at the end of the sample. I had problems with the initial
> >  > approach, when running test with multiple clients, the 1st client
> was
> >  > receiving all the messages. This behavior worked in topics because
> the
> >  > delivery semantics are different publish/subscribe vs point-to-
> point.
> >  >
> >  >
> >  > AIUI, queues are supposed to have a single reader and one or more
> >  > writers. Is that correct?
> >  >
> >  > If so, then using multiple samplers for the same queue would mean
> that
> >  > the queue must be handed off between samplers, so only one uses it
> at
> >  > a time.
> >  >
> >  > I'm not sure that it makes sense for a single queue to be shared
> >  > between multiple threads.
> >
> >
> > AIUI, you can have multiple writers and multiple readers but when a
> message is sent, only one reader will receive the message, in contrast
> with topics where all readers will receive it.
> 
> OK.
> 
> >  When having multiple threads I think nothing guarantees that the
> message sent/generated by the 1st thread will be received by the 1st
> thread (this should be guaranteed using correlation(?)) . However if
> the tests are sequential and no threads are used, only one receiver
> will exist at a given moment. This will work for my requirements
> because I am more focused on functional testing, i.e guarantee that a
> message is generated for a specific queue after some action is
> executed...
> 
> So one just has to be careful how the test plan is set up.
> 
> ==
> 
> It looks as if the same receive code could be shared for Topics and
> Queues.
> 
> Rather than using  QueueReceiver or TopicSubscriber, JMeter just needs
> to use the  MessageConsumer interface. Instances of which are created
> by using Session.createConsumer(Destination).

I tried those interfaces at start but didn't end too well, and sorry I can't remember why...
But go ahead try, I will test if you need.
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
> 



---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org


Mime
View raw message