https://bz.apache.org/bugzilla/show_bug.cgi?id=58506

Philippe Mouawad <[email protected]> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |NEEDINFO

--- Comment #4 from Philippe Mouawad <[email protected]> ---
(In reply to Gordon Daugherty from comment #2)
> This is a request for feedback to guide our development effort on this one.
> We're considering taking the approach given in my last comment:
> 
> When no "JNDI name Receive queue" is provided create a single temporary
> queue and act as if that temp queue's name was placed in the "JNDI name
> Receive queue" field. This will naturally cause receive timeouts to be
> applied AND will avoid the coordinated omission error in the metrics that
> JMeter collects.
> 
> Please respond by stating whether you have significant concerns about this
> approach. It'll keep the UI simple but will prevent JMeter from being able
> to generate a load using a limited number of threads each doing a blocking
> invocation.
> 
> If there are no objections to this approach we'll do it as-proposed to keep
> it simple. If you believe that users want to be able to generate load using
> a limited number of threads each doing a blocking invocation we can try to
> come up with an understandable way to present both options to users in the
> UI but I think it'll end up making the UI less understandable.

First thanks for your contribution.
Second, if possible, I think you should keep the current possibility to
generate load using a limited number of threads each doing a blocking
invocation, unless this is absolutely stupid, which for now is not sure in my
understanding.

What would be the impact on UI of this ?

-- 
You are receiving this mail because:
You are the assignee for the bug.

Reply via email to