Inter-cluster calls are getting dropped when someone picks up

Unanswered Question
Sep 27th, 2010
User Badges:

Inter-cluster calls are getting dropped when someone picks up.Its working fine with SIP phone but SCCP phones are dropping. Any idea ?

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Steven Holl Mon, 09/27/2010 - 06:08
User Badges:
  • Cisco Employee,

Sounds like a codec negotiation issue, perhaps?  Do you have CM traces from each node during an issue?

sanjayist Tue, 09/28/2010 - 00:35
User Badges:

Hi Steve,


Region settings looks fine.I made few test calls and attached traces.


Bad Call


Calling Party - 8-79-3007

Called Party - 8-124-5000

Time of call - 12:45 pm


Bad Call


Calling Party - 8-124-5000

Calling Party - 8-79-3007

Time of call - 12:47 pm


Good Call


Calling Party - 8-79-3003

Called Party - 8-124-5000

Time of call - 12:49 pm


Extension number -  3007 and 3003 are in AMD cluster which is facing the issue

Extension number - 5000 is registered with GGN cluster which is working fine.


Please suggest.

Felipe Garrido Tue, 09/28/2010 - 08:00
User Badges:
  • Cisco Employee,

I don't see a failed call at the times specified. Are you sure you collected the traces from the correct CUCM nodes? I do see a failed call between 3007 and 5000 at 15:37:40. That call fails due to a MTP allocation failure when trying to allocate "MTPTRANS."


09/28/2010 15:37:40.325 CCM|AgenaInterface(52)::sendErrorToMX, Audio Channel OLCAck with error recd


Can you verify what codecs that particular MTP supports? It looks as though the call is trying to use g729br8.


-Felipe

Steven Holl Tue, 09/28/2010 - 08:03
User Badges:
  • Cisco Employee,

I see the DA here:


09/28/2010 15:37:58.401 CCM|Digit analysis: analysis results|


I would expect to see an h225 setup out after this for an ICT call, especially since the phone is ringing, but that's not in there.  Can you collect the traces for all nodes in the cluster?  I think the ICT call went out a different node than the one the traces were collected for.  It looks like they're being sent off the sub?


I agree with Felipe that it looks like something regarding media or MTP allocation.  Shot in the wind, but perhaps a transcoder is being invoked for a g729 MTP call (which will always fail until 7.1.3 with universal transcoder support)?  Make sure the top MRG in the MRGL has only MTPs, and has an MTP that will support the type of codec being used across the trunk.


Message was edited by: Steven Holl | Saw Felipe's post.  Adding to my analysis based on that.

Actions

This Discussion