Need some help here. I've recently setup a ipip gw. The setup is PBX--> GW ---> IPIP GW --> CM. When the user from the PBX call the phones in CM, phones will ring but when the user pick up no audio can be heard from both sides.
When i did a debug voice rtp, i've got
*May 7 10:28:16.533: RTP(1668): fs tx s=IPIPGW(19358), d=IPPHONE(24
584), pt=18, ts=6E925A36, ssrc=13591C22
*May 7 10:28:16.557: RTP(1669): fs rx s=GW(16818), d=IPIPGW(1
8640), pt=18, ts=6E925AD6, ssrc=13591C22
*May 7 10:28:16.557: RTP(1669): fs tx s=IPIPGW(19358), d=IPPHONE(24
584), pt=18, ts=6E925AD6, ssrc=13591C22
Is it normal for the destination ip to be the ip phone's ip instead of callmanager. I've checked the mtp for the gateway configuration.
Any leads on this will be great.
Yes, it's normal for the destination address to be that of the phone. CCM does the call setup, but the audio flows between endpoints. It sounds like you have a codec mismatch. Make sure that you are using the same codec end to end or that you have a transcoder configured to handle the mismatch. Check the inbound/outbound dial peers on the GW and IPIPGW as well as the trunk/gateway configuration in CCM.
Hope this helps.
Without seeing your configuration, I am only making assumptions. First, are you doing Sip-To-Sip or Sip-to-H323, etc? Secondly, have you verified what codecs you are using on the different call legs. For instance, if the PBX to IP gateway is G729 but the call leg from the gateway to CUCM is G711 - you will need onboard transcoding resoures. If you require MTP, have you verified that MTP resources are available to the gateway? Are you using CUCM MTP or hardware resources on a router?
The MTP built-in to CUCM only supports G.711, are you using an IOS MTP?
Also, you don't have a voip dial-peer that will match inbound calls.
Why do you have an h323-gateway bind on the ipipgw? Is the ipipgw also terminating h323 traffic?
Do you have CUCM configured as an ICT or an H323 gw?
Two main reasons that rtp fails when the call setup works:
1. Codec mismatch - however, you will usually not get dead air but a disconnect.
2. Routing - Make sure that your phone subnet can talk to the IP addresses that you are tying your H323 to on the PBX gateway. CUCM may be able to talk to both but possibly the phone and gateway can't see each other.
Why would setup work? When CUCM can see the phones and also the gateway, setup will work, as it is in the middle. CUCM then tells the endpoints what IP/port they should communicate with for RTP - thus CUCM is no longer in the loop.
Thanks for the info. Is it neccessary for the endpoints to need to be able to communicate with the gateway and vice versa? Can we configure in such a way that the calls terminate at cm instead because i thought mtp with mtp checked it can do so. This way if we are to add other sites to the
You are correct that the endpoints need to be able to "see" where the RTP stream terminates. That can be the gateway or MTP. If you are wanting to use MTP, I would recommend offloading that work to a hardware based resource on a router, etc. Again, you also have to analyze codec compatibility with MTP.
Here is a snippet from the 6.X SRND:
An MTP can be used to transcode G.711 a-law audio packets to G.711 mu-law packets and vice versa, or it can be used to bridge two connections that utilize different packetization periods (different sample sizes). Note that re-packetization requires DSP resources in a CiscoÂ IOS MTP.
As was stated before, on server is only G711. To do G729 to G711, etc. you will require an IOS based MTP resource.
I've tried configuring the IOS mtp but it is not registering to the callmanager. My configurations are as below.
dsp services dspfarm
sccp local GigabitEthernet0/0
sccp ccm XX.XX.XX.XX identifier 1 version 4.1
sccp ccm group 1
associate ccm 1 priority 1
associate profile 1 register MTP001ebeea3e80
dspfarm profile 1 mtp
maximum sessions software 50
associate application SCCP
SCCP Admin State: UP
Gateway IP Address: XX.xx.xx.xx, Port Number: 2000
IP Precedence: 5
User Masked Codec list: None
Call Manager: xx.xx.xx.xx, Port Number: 2000
Priority: N/A, Version: 4.1, Identifier: 1
Software MTP Oper State: ACTIVE_IN_PROGRESS - Cause Code: TCP_CONN_ERROR
Active Call Manager: NONE
TCP Link Status: CONNECT_PENDING, Profile Identifier: 1
Reported Max Streams: 100, Reported Max OOS Streams: 0
Supported Codec: g729br8, Maximum Packetization Period: 60
Supported Codec: rfc2833 dtmf, Maximum Packetization Period: 30
What version of cucm are you running exactly?
Are you configuring this under mtp as an ios enhanced dspfarm?
Why are you restricting the software mtp to g729br8 rather than the more generic g729r8? The endpoints will negotiate the a or ab variant. The Cisco ip phones do not support g729br8.
Is there a router or firewall between the cucm and gw? They should both be on the same segment/vlan.
Finally, you should get in the habit of using a loopback address for the sccp/h323/sip binding rather than a physical interface.
I am running on Cm 4.1.3. I've configured it under as IOS enhanced MTP. There is no firewall in between but there is a gateway. Anyway the router with DSP and CM are not in the same segment but they are able to reach each other.
I have bind the loopback add as a h323 interface.
I'm having the same problem too. Is it CUCM Software MTP can support DTMF for T38? I have the same setup like you and all international and local call success but only fax t38 failed.
What is the result of 'show call act vo com' and 'show ip rtp conn' at the same time? The voice call is the signaling, the rtp conn is the voice packets, both are independent of each other. It is normal for the h323 signaling to be from CUCM and the rtp audio path directly from the ip phones, particularly if you have 'media flow-around' configured under 'voice service voip' on the cube.