asterisk users - Jul 2014 - Internal calls without voice transport

Hey,

we're experiencing a weird problem with Asterisk 1.8.13.1
(1:1.8.13.1~dfsg1-3+deb7). Calls that leave and enter Asterisk via
a PBX (sipgate.de) work perfectly fine, almost 100% of the time.

However, calls that are routed to sipgate.de, which then routes the
call back to our Asterisk instance are "silent" most of the time.

What I mean with that is that even though RTP traffic flows, neither
side can hear anything from the other.

This problem happens when people at site A dial someone at site
B using the number provided by sipgate.de, but also if people call
each other within a site through the external number, i.e. if I dial
089-1234567-100 from 089-1234567-200.

I have not been able to reproduce this problem with purely internal
calls, i.e. calling ext. 100 directly, so I am assuming there's
a problem due to sipgate's involvement. However, as far as
I understand, once the call is established (and both parties' phones
suggest that), the traffic flows only via Asterisk (directmedia
= update,nonat), so the problem is likely to be found there, no?

Before I shower you with debug logs and traces, I am wondering if
this sounds familiar to anyone??

Thanks,

-- 
martin | http://madduck.net/ | http://two.sentenc.es/
 
if god had meant for us to be naked,
we would have been born that way.
 
spamtraps: madduck.bogus at madduck.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: digital_signature_gpg.asc
Type: application/pgp-signature
Size: 1107 bytes
Desc: Digital signature (see
http://martin-krafft.net/gpg/sig-policy/999bbcc4/current)
URL:
<http://lists.digium.com/pipermail/asterisk-users/attachments/20140728/a7d0e89a/attachment.pgp>

By chance, I managed to fig into this a bit and found the exact
moment when audio stops. It is exactly the moment when the
counterparty picks up and RTP debug output says:

  Got  RTP packet from    46.244.255.146:8058 (type 00, seq 000680, ts
340914880, len 000160)
  Sent RTP packet to      46.244.255.146:8058 (type 00, seq 026000, ts
3578986600, len 000160)
      -- SIP/lehel-sipgate-00003573 answered SIP/lehel-martin-00003572
      -- Remotely bridging SIP/lehel-martin-00003572 and
SIP/lehel-sipgate-00003573
  Sent RTP P2P packet to 46.244.255.146:8058 (type 08, len 000160)
  Sent RTP P2P packet to 46.244.255.146:8058 (type 08, len 000160)

so RTP switches to RTP P2P and no more packets are received from the
phone.

I did have a sniffer running on 46.244.255.146, and Wireshark really
rocks, so now I know that the gateway firewall is at fault, and
indeed, for some reason, nf_conntrack_sip and nf_nat_sip were not
loaded. Now I am wondering how it worked in the first place, but
that's that. Maybe this will fix things.

Anyway, I don't quite yet understand what RTP P2P packets are or why
they are sometimes used and not at other times. I assume they are
packets intended to be exchanged directly between the two clients,
but since I have MixMonitor() on Asterisk, this shouldn't actually
be possible as Asterisk should always force itself into the middle.

Thoughts?

-- 
martin | http://madduck.net/ | http://two.sentenc.es/
 
dies ist eine manuell generierte email. sie beinhaltet
tippfehler und ist auch ohne gro?buchstaben g?ltig.
 
spamtraps: madduck.bogus at madduck.net
-------------- next part --------------
A non-text attachment was scrubbed...
Name: digital_signature_gpg.asc
Type: application/pgp-signature
Size: 1107 bytes
Desc: Digital signature (see
http://martin-krafft.net/gpg/sig-policy/999bbcc4/current)
URL:
<http://lists.digium.com/pipermail/asterisk-users/attachments/20140728/ec31f099/attachment.pgp>