[SR-Users] Kamailio 4.4.1 REGISTER Parse Error in PCSCF

[Ivailo Dobrev]

Hi Mihail,

There is something strange in your log:
parser/msg_parser.c:610]: parse_msg():  method:  
<*#001#020#002#021Žt#023ÄÀš±2Àš±d*REGISTER>(24)

For some reason bad SIP messages are generated somewhere. Maybe if you 
find the cause of this problem you will resolve the other.

Regards,
Ivo

On 05/30/2016 05:31 PM, Mihail Dakov wrote:
> Hi,
> Yes I got a different setup with kamailio 4.2.4 where I successfully
> register the same UEs. But there we got some kind of issue with the rx
> interface and before looking at the code we wanted to update kamailio
> version and then see if same rx issue happens. At the moment is not
> clear to me the rx issue therefore i couldn't describe it.
>
> On 05/30/2016 04:28 PM, Carsten Bock wrote:
>> Hi Mihail,
>>
>> you could also see the REGISTER in the PCAP, if you did not configure
>> your DNS/BIND properly...
>>
>> Thanks,
>> Carsten
>>
>> P.S.: I regularly test Zoiper with our own setup, it works... and it
>> even should work with the 4.4 version, you're running.
>>
>> 2016-05-30 17:25 GMT+03:00 Mihail Dakov <mihail.dakov at ng4t.com>:
>>>
>>> On 05/30/2016 04:11 PM, Carsten Bock wrote:
>>>> Hi,
>>>>
>>>> the Zone-File is the interesting part, not the Bind-Config.
>>>>
>>>> Did you configure the outbound Proxy on Zoiper properly?
>>> Yes, otherwise the REGISTER from the UE would be missing from the pcap,
>>> right?
>>>> You can actually see in the trace, that the Proxy-CSCF is forwarding
>>>> the packet to itself (looping) and at some point it drops the request.
>>>>
>>> I see that as well. I eliminated big part of the log as is the loop
>>> itself. After some time it gets stable. But to estimate the nature of
>>> the loop I get about 1M log lines with debug level 5 for 1 register
>>> procedure triggered from the UE.
>>>
>>> I guess i'll check with the trunk code meanwhile.
>>>> Thanks,
>>>> Carsten
>>>>
>>>> 2016-05-30 16:53 GMT+03:00 Mihail Dakov <mihail.dakov at ng4t.com>:
>>>>> Hi,
>>>>>
>>>>> DNS error? This is interesting. Well I got a local bind running with a
>>>>> dnszone provided from ims aio from kamailio.org site. You can have a
>>>>> look at the bind config which copied adjusting port numbers.
>>>>> I would say that it does work:
>>>>>
>>>>> user at host:etc/kamailio$ host hss
>>>>> hss.open-ims.test has address 192.168.177.100
>>>>> user at host:/etc/kamailio$ host icscf
>>>>> icscf.open-ims.test has address 192.168.177.100
>>>>> user at host:/etc/kamailio$ host pcscf
>>>>> pcscf.open-ims.test has address 192.168.177.100
>>>>> user at host:/etc/kamailio$ host scscf
>>>>> scscf.open-ims.test has address 192.168.177.100
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> On 05/30/2016 03:40 PM, Carsten Bock wrote:
>>>>>> Hi,
>>>>>>
>>>>>> looks to me, like a typical DNS error. How did you configure DNS?
>>>>>> You have BIND up and running and a proper Zone-File? Do the CSCF's
>>>>>> properly query the DNS?
>>>>>>
>>>>>> Thanks,
>>>>>> Carsten
>>>>>>
>>>>>> 2016-05-30 16:35 GMT+03:00 Mihail Dakov <mihail.dakov at ng4t.com>:
>>>>>>> Hi Folks,
>>>>>>>
>>>>>>> I have the strangest problem just after installing latest stable
>>>>>>> version. I had tried both ways using the repo
>>>>>>> (http://deb.kamailio.org/kamailio44) and downloading code with local
>>>>>>> compile (https://www.kamailio.org/w/2016/05/kamailio-v4-4-1-released/)
>>>>>>>
>>>>>>> The initial scenario is to get M2M with commercial IP and subsequently
>>>>>>> use the Rx interface to do the same. We are blocked because we cannot
>>>>>>> even achieve clean register for the UEs.
>>>>>>>
>>>>>>> UE1(galaxy s4 zoiper) -------- Kamailio IMS aio
>>>>>>>                                 |
>>>>>>> UE2(galaxy s4 zoiper) -------- |
>>>>>>>
>>>>>>>
>>>>>>> What happens is that the first register message triggers some
>>>>>>> quasi-endless loop in the PCSCF. Hence it cannot be forwarded to SCSCF,
>>>>>>> ICSCF and HSS. But this doesn't always happen and between all the parse
>>>>>>> problems there are some register messages which get passed to the scscf
>>>>>>> icscf and hss. Somehow the parser feeds itself from the previous parse
>>>>>>> run and thus loops quite intensively. The message I get is SIP/2.0 400
>>>>>>> CSeq method does not match request method.
>>>>>>> I must say that reiterating the registration process sometimes i get the
>>>>>>> register and sometimes i don't.
>>>>>>>
>>>>>>> Has anyone encounter similar issue for similar use case and can share
>>>>>>> some thoughs about a fix?
>>>>>>>
>>>>>>> I attached relevant log file, pcaps and configuration files. Let me know
>>>>>>> if more config files are needed.
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list
>>>>>>> sr-users at lists.sip-router.org
>>>>>>> http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users
>>>>>>>
>>>>>>
>>>>>>
>>>>> _______________________________________________
>>>>> SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list
>>>>> sr-users at lists.sip-router.org
>>>>> http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users
>>>>>
>>>>
>>>>
>>>
>>> _______________________________________________
>>> SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list
>>> sr-users at lists.sip-router.org
>>> http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users
>>>
>>
>>
>
>
> _______________________________________________
> SIP Express Router (SER) and Kamailio (OpenSER) - sr-users mailing list
> sr-users at lists.sip-router.org
> http://lists.sip-router.org/cgi-bin/mailman/listinfo/sr-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sip-router.org/pipermail/sr-users/attachments/20160530/dca1dd11/attachment.html>