CUCM 7.1.5 SIP trunk and MTP issue

[kyiver_voip]

Hello,

I am trying to set up a SIP trunk for a client between a recently provisioned CUCM 7.1.5 server and a VoIP service provider. The software MTP is provisioned and G711alaw codec is set up on both sides. The phone and MTP are also in the same Region configured for G711. However I get the following error in the syslog on outgoing calls:

CCM_CALLMANAGER-CALLMANAGER-3-MtpNoMoreResourcesAvailable: No more MTP resources available.

The incoming calls do not go through at all, but the error dos not show up in the syslog. I have also tried to set up the SIP trunk between new CUCM 7.1.5 and the old CCM 4.1.3 and there is the same issue.

I have been experimenting with different settings for a few days now, but everything is useless. This is becoming very frustrating. Please help. All suggestions are welcome!

Thanks.

Andri

[dkemp]

Andri,

     On your SIP trunk do you have a Media Resource Group List configured? Is the MRGL configured with the correct group with the MTP in it as a resource?

Cheers

Dave

[kyiver_voip]

Hi Dave, thanks for your reply.

Yes the MRGL is configured with the software MTP and assigned to the trunk. In RTMP it's showing 24 recources available.

I went through the standard check list many times but can't figure out the problem...

Thanks.

Andri

[Jing Hua]

Did you try to restart "Cisco IP Voice Media Streaming App " service? For your inbound call from service provider you might want to check what digits service provider sends in and if you have any digits manipulation done after call hit ccm. Pls also make sure CSS of your sip trunk can reach your ip phones.

[dkemp]

Andri,

Also give the trunk and mtp a reset, if still no good then will need to see a cm trace to identify the issue.

Cheers

Dave

[kyiver_voip]

Thanks everyone.

I have restarted the services as well as the server itself - no effect.

Once again, it's software MTP, not cisco IOS. And MTP required is checked as per my original post. I tried to uncheck it but then the calls fail immediately. But I believe we need MTP anyway.

I have attached here the SDL log. The first call there is outgoing, the second one is incoming. If anyone could help me to analyse the trace that would be great. There is also a screen capture attached which is showing what options for the trace are enabled. Let me know if I need to enable anything else.

Looking at this trace I noticed that some lines are referring to node 'SIG-UCM01" which was the original name before I renamed it to the IP address. So I just changed it back to the host name back again and restarted the server, - zero effect.

Please note that it's a new installation (yet the license is already installed) and I can changed all settings as much as needed.

Cheers,

Andri

[cjoseph23]

I am assuming you are using CUBE as the demarcation between the Service provider and your CUCM?  IS the Topology as follows:

CUCM ----sip -----CUBE ----sip ----Service Provider SBC

If so then can you run:

debug ccsip messages

Place an outbound and inbound call and see what errors you are getting.

regards,cj

[kyiver_voip]

CJ,

No, there is no CUBE, but the CUCM is NATed to a dedicated public IP address. I believe if the MTP is enabled the callmanager should act as a signgle point for the media relay? What would be the advantages of using CUBE?

Also, as I mentioned before the same problem happen even when I setup a SIP trunk between two CUCM clusters in our network - CUCM 7.1.5 and CCM 4.1.3.

Thanks.

[cjoseph23]

CUBE will provide a demarcation point between your CUCM and the service provider.  Much the same as a traditional PSTN gateway.  I have not done a SIP deployement withough using CUBE (or Acme Packet) so I am not sure what config would need to look like on the CM.

The troubleshooting I was asking for was more for the inbound call leg from the provider.  Since you are not seeing that in your SDL traces I thought we could capture that on the CUBE.

I see that someone found some of your phones using g722 codec (probably 7942/62 phones).  I am guessing you don't have HW transcoding resources.

Sorry i couldn't be of more assistance.

cj

[htanay]

Hi,

Do you have "mtp required" checked on the trunk ?

Although that might be required for midcal signalling....

Try unchecking that.

Check if the MTP resources allocated in the MRG/MRGL are registered (and if these are IOS mtp, check if the router ios is compatible with CUCM: e.g. sccp ccm 192.168.12.44 identifier 1  version 6.0 alt show run | sec sccp)

Thanks,

[Jing Hua]

Hi,

I've had a look at the SDL you uploaded and looks like cucm was trying to use "MTP_2" but got '

No more MTP resources available' error .If we have SDI trace then it might be easier to find out why. Coulple of things we need:

1. calling number and dialed digits and test time.

2. detailed ccm sdi and sdl logs cover the test. Please refer link below for trace setting (if you can tick 'sip stack' for sdi setting the it would be better):

http://www.cisco.com/en/US/partner/products/sw/voicesw/ps556/products_tech_note09186a0080094e89.shtml

note:

Please make sure you restart service on "MTP_2" server not others

[kyiver_voip]

Hi Jing,

I have uploaded here more traces as you asked. These are detailed SDI and SDL traces for CM and Cisco IP Voice Media Streaming application. SIP stack box is checked too. It includes too calls:

11:50AM - outgoing from 5001 to 07500932480

11:51AM - incoming from +447500932480 to 02033880063 (hunt group includes 5001 and 5000)

There is only one server at the moment. I am not sure why the MTP recource has a name MTP_2. It was automatically assigned and I have no idea why...

Thanks!

Jing Hua wrote:
Hi,
I've had a look at the SDL you uploaded and looks like cucm was trying to use "MTP_2" but got '
No more MTP resources available' error .If we have SDI trace then it might be easier to find out why. Coulple of things we need:
1. calling number and dialed digits and test time.
2. detailed ccm sdi and sdl logs cover the test. Please refer link below for trace setting (if you can tick 'sip stack' for sdi setting the it would be better):
http://www.cisco.com/en/US/partner/products/sw/voicesw/ps556/products_tech_note09186a0080094e89.shtml
note:
 Please make sure you restart service on "MTP_2" server not others  

[Jing Hua]

Went through sdi and found allocate MTP failed because 1 device is using g711u but another is using g722(most likely phone).  Please disable g722 below and try again:

1. ccm admin > system > service parameters > Cisco Callmanager service  > "G722 Codec Enabled" > disabled

2. ccm admin > system > enterprise parameters > Advertise G.722 Codec > disabled

provider sent in digits 02033880063 but there is no such a number in CCM that calling search space "Internal:Emergency" can reach. Please double check your settings in ccm to make sure there is a number 02033880063 that trunk "pstn_sip_trunk" can reach.

[Jing Hua]

go to bed now and hopefully can hear some good news in the morning

[kyiver_voip]

Hi Jing, thanks for your instructions. Sorry for the delay I have been having another urgent issue here, so couldn't test properly.

But I applied those changes as you advised. Now when I call from a desk phone (7960) I get the outgoing call connected but without media, and it gets disconnected after about 10 sec. I suspect this can be the firewall issuu, so I will talk to the administrator.

However, when calling from IP coomunicator on my laptop, the issue remain the same. So when I call from IP Communicator the call gets disconnected wihin moment when calls reach my mobile phone, - same as prior to applying your instructions. I have even specifically disabled the G722 for this account on CUCM with no effect at all. Do you have any ideas why this can happen?

Also the situation with incoming calls hasn't changed. Even though I am sure  that the Hunt Pilot is in the Internal partition. I can call the hun pilot number internally.

Many thanks.

Andri