Re: [Opalvoip-devel] slow response to SIP OPTIONS request by client

[Dan J. <dan...@gm...>]

Hi, Robert,

I've traced the slowness I'm experiencing to be caused by slow STUN
responses.
I've noticed that sometimes the reply to the OPTIONS is sent only after a
STUN binding message is sent and a response is received.

Can you explain why the client might need to send the STUN messages?

Thanks,
Dan


On Wed, Jul 29, 2009 at 9:21 AM, Dan Julius <dan...@gm...> wrote:

> Thanks Robert.
> I think this only applies to the OPTIONS message.
>
> I'll keep you posted if I find anything.
>
> Dan
>
> On Wed, Jul 29, 2009 at 4:38 AM, Robert Jongbloed <
> ro...@vo...> wrote:
>
>>  That is truly bizarre. I assume your multiple “instances” are completely
>> separate processes? I cannot imagine how they could be affecting each other
>> over a process boundary!
>>
>>
>>
>> Is it only the OPTIONS that is slow? How about INVITE’s?
>>
>>
>>
>>
>>
>> A quick look at the code and there is just no significant code between the
>> transmission of the 100 Trying (sipep.cxx line 515) and the 200 OK
>> (sipep.cxx line 941). There is nothing I can see that would block on a mutex
>> or I/O. There is potentially a DNS lookup in there, but why that would cause
>> a 2 second delay I have no clue, that tends to be instantaneous or 30 odd
>> seconds.
>>
>>
>>
>> The only other thing is running out of CPU, but only 10 instances does not
>> seem near enough for any modern processor. Especially if you do not happen
>> to have any calls going at the time.
>>
>>
>>
>>
>>
>> I am afraid I think you are on your own on this one!
>>
>>
>>
>> Robert Jongbloed
>>
>> OPAL/OpenH323/PTLib Architect and Co-founder.
>>
>>
>>
>> *From:* Dan Julius [mailto:dan...@gm...]
>> *Sent:* Wednesday, 29 July 2009 2:25 AM
>> *To:* opa...@li...
>> *Subject:* [Opalvoip-devel] slow response to SIP OPTIONS request by
>> client
>>
>>
>>
>> Hi,
>>
>> I'm using opal 3.6.3.
>> I've noticed strange behavior when running multiple instances of my client
>> on the same machine (windows server 2003), all of them registered with a
>> single Asterisk server.
>> I have asterisk configured to periodically send SIP OPTIONS messages to
>> each of the clients.
>>
>> When I have more than 10 clients running there is always one instance that
>> responds very slowly to the SIP OPTIONS -
>> I see that after receiving the message, a TRYING message is sent
>> immediately back to the server, but then it takes over  2s till the OK
>> message is sent (measured with wireshark).
>> Moreover, if I restart this instance, a different instance becomes slow.
>>
>> How are these instances related in ptlib/opal (there is no relation in my
>> code that I am aware of).
>> Any ideas what could be causing this slowness?
>>
>> Thanks,
>> Dan
>>
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Let Crystal Reports handle the reporting - Free Crystal Reports 2008
>> 30-Day
>> trial. Simplify your report design, integration and deployment - and focus
>> on
>> what you do best, core application coding. Discover what's new with
>> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
>> _______________________________________________
>> Opalvoip-devel mailing list
>> Opa...@li...
>> https://lists.sourceforge.net/lists/listinfo/opalvoip-devel
>>
>>
>