Possible bug: stateless behavior when transport=tcp

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



hi!

I would say naturally (if it's possible) to keep the same transport for a
dialog. looks like you have two dialogs in your example.
Let look your case when there are two SIP proxies behind load-balancer.
What do you expect in this case?

On Tue, Mar 17, 2015 at 9:39 AM, Harald Radke <harryrat at gmx.de> wrote:

> hmm....ok, let me get this  straight:
>
> two siphones, ALICE and BOB, and one Asterisk server
>
> Alice's sip phone has as listener port say 10000,
> Bob's sip phone has as  listener port 11000,
> Asterisk sits on 5060
>
> Asterisk acts as registrar and proxy for Alice and as proxy for Bob.
>
> Alice registers at Asterisk: (only ports shown, no IPs)
> random port ( say 22222) --> 5060  REGISTER
> 22222 <----- 5060  OK
> ...
> Alice places as a call:
> 22222 -------> 5060 INVITE xyz at abc
> 22222 <------  5060  100 Trying/180 Ringing /200 Ok
> ...
> Now bob calls Alice, via Asterisk:
> Bob
> random port (say 33333) -----> 5060 INVITE alice at ....
>
> and now Asterisk does:
> random port (say 44444) ------> Alice 22222
>
> ? Is that correct ?
>
> I cannot test this right now, since I have no Asterisk nor a configured
> proxy/registrar.
>
> What I tested was:
>
> Alice 10000
> Boc  11000
>
> Bob calls Alice:
> random (22222) -----> 10000 INVITE
> 22222 <--------- 10000 OK
>
> Alice sends an OPTIONS request:
> random (33333) -----> 11000 OPTIONS
> 33333 <-------- 11000 OK
>
> So basically Alice creates for her request a new TCP connection to Bob,
> not using the existing one
>
> -> Requests in different directions between the same hosts results in
> different TCP connections
>
> Why now in your case that does not seem to be the case...I dont know (-:
> it would be good to test another Proxy/Registrar based
> on PJSIP
>
> Regards,
>
> Harry
>
>
>
>  *Gesendet:* Dienstag, 17. M?rz 2015 um 08:24 Uhr
> *Von:* Anatoli <me at anatoli.ws>
> *An:* "'pjsip list'" <pjsip at lists.pjsip.org>
> *Betreff:* [pjsip] Possible bug: stateless behavior when transport=tcp
> Hi all,
>
> It looks like there is a bug in PJSIP, but I've submitted a bug in Asterisk
> tracker as I'm not sure if it occurs only in connection with res_pjsip or
> as
> stand-alone as well.
>
> Basically, when an endpoint registers via TCP, PJSIP processes everything
> as
> corresponds: the communication is via TCP, REGISTER/ACK pass OK, the
> endpoint can even establish calls. But, when another endpoint calls to the
> first one, PJSIP tries to send INVITE message in a new TCP session
> establishing it to the port already opened for the first session. This
> would
> be OK for a UDP, but not for TCP. Please let me know if you observe the
> same
> behavior without asterisk as well.
>
> For more information please check the bug report here:
> https://issues.asterisk.org/jira/browse/ASTERISK-24889.
>
> Regards,
> Anatoli
>
>
> _______________________________________________
> Visit our blog: http://blog.pjsip.org
>
> pjsip mailing list
> pjsip at lists.pjsip.org
> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
>
> _______________________________________________
> Visit our blog: http://blog.pjsip.org
>
> pjsip mailing list
> pjsip at lists.pjsip.org
> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
>
>


-- 
Best Regards,
Serhiy Serbin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.pjsip.org/pipermail/pjsip_lists.pjsip.org/attachments/20150317/e810982a/attachment.html>


[Index of Archives]     [Asterisk Users]     [Asterisk App Development]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux]     [Linux OMAP]     [Linux MIPS]     [Linux API]
  Powered by Linux