[OpenSIPS-Users] 408 Request Timeout with UDP

Bogdan-Andrei Iancu bogdan at opensips.org
Fri Aug 7 12:21:41 CEST 2015


Hi Nabeel,

Using TCP gives an advantage over UDP as being connection oriented, it 
force the NAT pinhole to stay open. Nevertheless, using the SIP pinging 
with UDP should also have fixed the problem.

I will update the default cfg to use SIP pinging rather than simple UDP 
pinging.

Thanks and Regards,

Bogdan-Andrei Iancu
OpenSIPS Founder and Developer
http://www.opensips-solutions.com

On 07.08.2015 12:32, Nabeel wrote:
> SamyGo,
>
> I tried the test you suggested but it did not work.
>
> Bogdan,
>
> After switching to TCP and adding your suggestions along with the 
> following parameters, the timeout errors seem to be resolved:
>
> *tcp_async=1*
> *tcp_connect_timeout=99999*
> *tcp_send_timeout=99999*
> *
> modparam("nathelper", "nortpproxy_str", "a=nortpproxy:yes\r\n")
>
> *
> I'm not sure in exactly what combination works best, but perhaps these 
> should be included in the default residential script?
> *
> *
> Thanks for the help... I'll be back with more questions.
>
>
> On 6 August 2015 at 15:41, Bogdan-Andrei Iancu <bogdan at opensips.org 
> <mailto:bogdan at opensips.org>> wrote:
>
>     Hi Nabeel,
>
>     This time the SIP trace looks ok - the message is sent to a public
>     IP and not a private one (as shown in the prev capture). Also the
>     usrloc data looks ok, telling that the NAT traversal works ok.
>     As you have Mobile Data, some operators do filter SIP, so not sure
>     what say.
>
>     Do you see the sip messages going back and forward during the
>     registration process ?
>
>     An issue may be the fact the default script does non-SIP pinging
>     (which is unidirectional), so the NAT may close. Add the followings:
>     1) modparam("nathelper", "sipping_bflag", "SIP_PING_FLAG")
>     modparam("nathelper", "sipping_from", "sip:pinger at xx.xx.xx.xx"
>     <mailto:sip:pinger at xx.xx.xx.xx>)
>
>     2) setbflag(SIP_PING_FLAG); before doing save()
>
>     You should see OpenSIPS doing keep alive with OPTIONS requests.
>     Regards,
>
>     Bogdan-Andrei Iancu
>     OpenSIPS Founder and Developer
>     http://www.opensips-solutions.com
>
>     On 06.08.2015 14:59, Nabeel wrote:
>>
>>     My OpenSIPS runs on a public IP.  The callee was connected to
>>     Wi-Fi in my first test earlier, but in the second test the callee
>>     was connected to a public IP  (public mobile network).  In both
>>     cases, the same '404 timeout' error occurred on call attempt. 
>>     The SIP trace for the second case is at this link:
>>
>>     http://pastebin.com/jGxRQ34q
>>
>>     Regarding private IP, you said it's impossible to route from
>>     public IP to private IP.  Although at the IP level this may be
>>     true, even if the user is on Wi-Fi, the whole point of NAT
>>     traversal is that the user's public IP is discovered and the call
>>     can get connected, is that not right?  I'm fact,  using a TURN
>>     server and a different SIP proxy, I was able to connect these
>>     same devices under the same networks, so I know this should be
>>     possible.  I feel something is not configured correctly in
>>     OpenSIPS / rtpproxy.
>>
>>     I did "opensipsctl ul show" and the results seem normal; please
>>     check it:
>>
>>     http://pastebin.com/n1BbTuMK
>>
>>     Perhaps the NAT processing just needs a bit more time; in thar
>>     case what are the config options to increase the request timeout
>>     for UDP?  I have seen the 'tcp_send_timeout' and
>>     'tcp_connect_timeout' options for TCP, but please let me know if
>>     there are similar options for UDP.
>>
>>     On 6 Aug 2015 12:08, "Bogdan-Andrei Iancu" <bogdan at opensips.org
>>     <mailto:bogdan at opensips.org>> wrote:
>>
>>         Nabeel,
>>
>>         I suppose you OpenSIPS seats on a public IP, right ? The
>>         callee looks to have a private IP. And, at IP level, it is
>>         impossible to route from a public IP to a private one.
>>
>>         I see your script has NAT traversal support. My question is -
>>         did the callee properly registered via this script ? can you
>>         do an "opensipsctl ul show" to see the callee's registration ?
>>
>>         Regards,
>>
>>         Bogdan-Andrei Iancu
>>         OpenSIPS Founder and Developer
>>         http://www.opensips-solutions.com
>>
>>         On 06.08.2015 07:14, Nabeel wrote:
>>>         Hi,
>>>
>>>         Yes, the destination IP is 192.168.0.19:60912
>>>         <http://192.168.0.19:60912/> and both phones are registered
>>>         to OpenSIPS.  In this case, the callee is connected to Wi-Fi
>>>         (hence 192.xx IP address) and the caller is connected to a
>>>         mobile network.
>>>
>>>         The opensips.cfg I am using was generated from 'make
>>>         menuconfig', except with the addition of "alias=domain.com
>>>         <http://domain.com>". I have attached my config file at this
>>>         link:
>>>
>>>         http://pastebin.com/0QRyC938
>>>
>>>
>>>
>>>         On 6 August 2015 at 05:00, SamyGo <govoiper at gmail.com
>>>         <mailto:govoiper at gmail.com>> wrote:
>>>
>>>             Hi Nabeel,
>>>             Quick question; what is this destination ip?
>>>             192.168.0.19:60912 <http://192.168.0.19:60912> ? -
>>>             Destination User Agent Registered on OpenSIPS?
>>>             Can you share the opensips.cfg code snippet for this call ?
>>>
>>>             On Wed, Aug 5, 2015 at 11:55 PM, Nabeel
>>>             <nabeelshikder at gmail.com
>>>             <mailto:nabeelshikder at gmail.com>> wrote:
>>>
>>>                 Hi,
>>>
>>>                 I am using the residential script generated by 'make
>>>                 menuconfig', with UDP and NAT support enabled.  I
>>>                 added "alias=domain.com <http://domain.com>" to the
>>>                 config because otherwise the UA did not register
>>>                 with my domain (username at domain.com
>>>                 <mailto:username at domain.com>). When I attempt to
>>>                 make a call, I see '408 Request Timeout' in the sip
>>>                 trace and the call does not connect.  Please check
>>>                 the log/trace below and advise how to fix this.
>>>
>>>                 SIP trace:
>>>
>>>                 http://pastebin.com/u5h9qGNr
>>>
>>>                 OpenSIPS log:
>>>
>>>                 http://pastebin.com/B8PUCKh0
>>>
>>>                 _______________________________________________
>>>                 Users mailing list
>>>                 Users at lists.opensips.org
>>>                 <mailto:Users at lists.opensips.org>
>>>                 http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>>
>>>
>>>
>>>             _______________________________________________
>>>             Users mailing list
>>>             Users at lists.opensips.org <mailto:Users at lists.opensips.org>
>>>             http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>>
>>>
>>>
>>>
>>>         _______________________________________________
>>>         Users mailing list
>>>         Users at lists.opensips.org  <mailto:Users at lists.opensips.org>
>>>         http://lists.opensips.org/cgi-bin/mailman/listinfo/users
>>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opensips.org/pipermail/users/attachments/20150807/89e06d49/attachment-0001.htm>


More information about the Users mailing list