Hi Robert,

I finally found the main issue with my deployment - MTU value on signalling
interfaces.
While main ifaces (eth0) was correctly set up by DHCP, signaling ifaces
used default value (1500), which caused network issues and strange
failures. I was able to make a call using Zoiper (Jitsi doesn't work for
some reason).

Thank you for help.

On Wed, Apr 20, 2016 at 4:58 PM, Sergey Kolekonov <[email protected]>
wrote:

> Just an update. Looks like I hit this problem (MTU issue) [0].
> After disabling some encodings, I moved forward, but still no luck - I'm
> getting 503.
>
> This is what I see in Sprout logs:
>
>> 20-04-2016 13:43:46.932 UTC Verbose pjsip: tcpc0x7f588c5c TCP transport
>> 192.168.1.5:33632 is connected to 192.168.1.6:5058
>> 20-04-2016 13:43:50.003 UTC Verbose pjsip: tcpc0x7f588c5c TCP connection
>> closed
>> 20-04-2016 13:43:50.003 UTC Debug pjsip: tsx0x7f588c5c5 State changed
>> from Calling to Terminated, event=TRANSPORT_ERROR
>> 20-04-2016 13:43:50.003 UTC Debug basicproxy.cpp:213: tsx0x7f588c5c5548 -
>> tu_on_tsx_state UAC, TSX_STATE TRANSPORT_ERROR state=Terminated
>
>
> This seems to be a reason for 503 errors.
>
> [0] http://lists.jitsi.org/pipermail/users/2013-December/005742.html
>
> On Wed, Apr 20, 2016 at 2:11 PM, Sergey Kolekonov <[email protected]
> > wrote:
>
>> Hi Robert,
>>
>> I've redeployed my environment and now I can see INVITE messages with
>> appropriate User-Agent, but the error is still here.
>> You can find a record at ./var/log/bono/bono_20160420T100000Z.txt line
>> 14364
>>
>> Can you please advice something?
>> Thanks.
>>
>> On Tue, Apr 19, 2016 at 9:27 PM, Robert Day (projectclearwater.org) <
>> [email protected]> wrote:
>>
>>> Hi Sergey,
>>>
>>>
>>>
>>> The Ralf logs shouldn’t cause call failures – Ralf is a component for
>>> enabling offline billing (also known as Rf billing), and those logs just
>>> indicate that you can’t connect to a Charging Collection Function at
>>> example.com.
>>>
>>>
>>>
>>> I can’t see any INVITE messages with “User-Agent: Jitsi” in the Bono
>>> logs you’ve sent, only REGISTER, SUBSCRIBE and OPTIONS messages. It’s worth
>>> checking local network capture with Wireshark or a similar tool when you
>>> make calls with Jitsi, in case it’s just sending messages to the wrong
>>> place, or resolving example.com incorrectly, and reporting a 408
>>> timeout locally.
>>>
>>>
>>>
>>> Best,
>>>
>>> Rob
>>>
>>>
>>>
>>>
>>>
>>> *From:* Sergey Kolekonov [mailto:[email protected]]
>>> *Sent:* 19 April 2016 13:46
>>>
>>> *To:* Robert Day (projectclearwater.org) <[email protected]>
>>> *Cc:* [email protected]
>>> *Subject:* Re: [Project Clearwater] [Clearwater] Sprout routing problem
>>>
>>>
>>>
>>> Robert,
>>>
>>>
>>>
>>> Adding alias_list helped. I also added scscf.sprout.example.com to DNS
>>> as it wasn't here and caused failures.
>>>
>>> After these steps some live tests started to pass (Call Barring, for
>>> example). But I'm still unable to make a call using real clients,
>>>
>>> as there're some 408 errors. Ralf logs also look suspicious:
>>>
>>> 19-04-2016 12:23:23.508 UTC Error diameterstack.cpp:293: Routing error:
>>> 'No remaining suitable candidate to route the message to' for message with
>>> Command-Code 271, Destination-Host  and Destination-Realm example.com
>>>
>>>
>>>
>>> I've attached full logs of Ralf, Sprout and Bono.
>>>
>>>
>>>
>>> Thanks for help!
>>>
>>>
>>>
>>> On Tue, Apr 19, 2016 at 12:20 AM, Robert Day (projectclearwater.org) <
>>> [email protected]> wrote:
>>>
>>> Hi Sergey,
>>>
>>>
>>>
>>> Can you try adding ‘alias_list=”sprout.example.com”` to
>>> /etc/clearwater/shared_config and running “sudo service sprout stop” to
>>> restart Sprout? That’s the domain name in the Route header of that
>>> REGISTER, and adding that will make sure that it’s recognised as local
>>> (which might not be happening in this case, and so the REGISTER isn’t being
>>> processed properly).
>>>
>>>
>>>
>>> If that doesn’t help, can you send across the full Sprout log – for
>>> example, by attaching it to an email (compressing it if it’s large)? That
>>> will help me see everything that’s going on with the REGISTER, from the
>>> point it enters the system onwards, and should also help me see the initial
>>> configuration of the system (there are some logs starting with “Local host
>>> aliases:” which will be useful for figuring this out)
>>>
>>>
>>>
>>> Best,
>>>
>>> Rob
>>>
>>>
>>>
>>> *From:* Sergey Kolekonov [mailto:[email protected]]
>>> *Sent:* 18 April 2016 17:05
>>> *To:* Robert Day (projectclearwater.org) <[email protected]>
>>> *Cc:* [email protected]
>>> *Subject:* Re: [Project Clearwater] [Clearwater] Sprout routing problem
>>>
>>>
>>>
>>> Hi Robert,
>>>
>>>
>>>
>>> Thanks a lot for the fast response.
>>>
>>> This solved my problem, but I ran into another [0].
>>>
>>> Now Sprout can not get subscriber data. The problem seems to be on
>>> Homestead side [1]:
>>>
>>> looks like it receives a request without necessary fields.
>>>
>>>
>>>
>>>
>>>
>>> [0] http://paste.openstack.org/show/494445/
>>>
>>> [1] http://paste.openstack.org/show/494449/
>>>
>>>
>>>
>>> On Mon, Apr 18, 2016 at 5:22 PM, Robert Day (projectclearwater.org) <
>>> [email protected]> wrote:
>>>
>>> Hi Sergey,
>>>
>>>
>>>
>>> From the log you provided, it looks like the REGISTER is being processed
>>> by the BGCF (which routes calls outside the deployment), not the S-CSCF
>>> (which processes registrations). We fixed an issue recently where the BGCF
>>> and S-CSCF were both on port 5054 and so sometimes requests would go to the
>>> wrong one (
>>> https://github.com/Metaswitch/sprout/commit/efd13e91a3e2a16d92b159aa9ec73faf25257935#diff-75af8e842ec03e21817a448de6bdf026R208
>>> ).
>>>
>>>
>>>
>>> Could you try setting “bgcf=0” (to disable the BGCF) in
>>> /etc/clearwater/shared_config, and running “sudo service sprout stop” to
>>> restart Sprout? Hopefully that will resolve your issue.
>>>
>>>
>>>
>>> By the way, you’re not currently subscribed to the mailing list, which
>>> means your posts have to be approved by an administrator before anyone sees
>>> them – you can sign up at
>>> http://lists.projectclearwater.org/mailman/listinfo/clearwater_lists.projectclearwater.org
>>> and avoid that step.
>>>
>>>
>>>
>>> Best,
>>>
>>> Rob
>>>
>>>
>>>
>>>
>>>
>>> *From:* Clearwater [mailto:
>>> [email protected]] *On Behalf Of *Sergey
>>> Kolekonov
>>> *Sent:* 18 April 2016 14:12
>>> *To:* [email protected]
>>> *Subject:* [Project Clearwater] [Clearwater] Sprout routing problem
>>>
>>>
>>>
>>> Hello,
>>>
>>>
>>>
>>> I need some help with Sprout issue.
>>>
>>> I've deployed Clearwater using clearwater-heat [0] on top of OpenStack
>>> Mitaka.
>>>
>>>
>>>
>>> So the problem is that I'm unable to register/make a call.
>>>
>>> After some investigations I found the following logs from Sprout [1].
>>>
>>> In short, it answers SIP/2.0 404 No route to target when a client tries
>>> to register.
>>>
>>> Results are equal when I use both the real client (Jitsi)
>>> and clearwater-live-test [2].
>>>
>>>
>>>
>>> Could you please help to understand what can be wrong?
>>>
>>>
>>>
>>> Thank you.
>>>
>>>
>>>
>>> [0] https://github.com/Metaswitch/clearwater-heat
>>>
>>> [1] http://paste.openstack.org/show/494412/
>>>
>>> [2] https://github.com/Metaswitch/clearwater-live-test
>>>
>>>
>>>
>>> --
>>>
>>> Regards,
>>>
>>> Sergey Kolekonov
>>>
>>>
>>>
>>>
>>>
>>> --
>>>
>>> Regards,
>>>
>>> Sergey Kolekonov
>>>
>>>
>>>
>>>
>>>
>>> --
>>>
>>> Regards,
>>>
>>> Sergey Kolekonov
>>>
>>
>>
>>
>> --
>> Regards,
>> Sergey Kolekonov
>>
>
>
>
> --
> Regards,
> Sergey Kolekonov
>



-- 
Regards,
Sergey Kolekonov
_______________________________________________
Clearwater mailing list
[email protected]
http://lists.projectclearwater.org/mailman/listinfo/clearwater_lists.projectclearwater.org

Reply via email to