On 04/29/2011 11:16 PM, Worley, Dale R (Dale) wrote:
> ________________________________________
> From: [email protected] 
> [[email protected]] On Behalf Of Joegen Baclor 
> [[email protected]]
>
> So I think the safest move (might not be the most elegant) is
> to preserve the status quo and just get rid of the nasty queue is full
> error message.
> _______________________________________________
>
> I hope you don't mean to prevent queue full situations from being logged.  
> There is a *reason* we put
> that message in:  If a queue fills and chokes the process so it doesn't do 
> any useful work, then there
> is some way to determine *which* queue filled, and thus, which thread has 
> gotten blocked, so we have
> some hope of finding and solving the problem.
>
> Dale
Hi Dale,

Thanks for chiming in.  I certainly did not mean "to prevent queue full 
situations from being logged".  I apologize for being vague.  What I 
mean is to get rid of the bug that causes that piece of log to appear 
when SipTcpClient objects are destroyed.  The queue that is being filled 
up is the SipTcpServer queue which is not doing anything in the run 
method except wait for the exit event.  In fact, SipTcpServer is not 
even start()'ed prior to the patch.  This has been a long standing bug 
but is considered benign and is one of the items pointed out in this 
task (http://track.sipfoundry.org/browse/XX-8063).
_______________________________________________
sipx-dev mailing list
[email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-dev/

Reply via email to