PRI MGCP behavioral question

Unanswered Question
Nov 30th, 2007
User Badges:

I have remote sites with local PRI's for dialtone connected to centralized CCM over an MPLS cloud.

On occasion, the wan will hiccup and go down for a few seconds. During this time, the PRI will fail as well. Most times, the PRI will re-register when the WAN returns to service, but some times I will have to do a manual reset to correct this.

I am running SRST at these sites, but the outage is not long enough to trigger SRST.

My question is this: Is this expected behaviour? Also, can anyone clarify how the MGCP process works with WAN failure.


  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 3 (1 ratings)
balitewiczp Fri, 11/30/2007 - 07:56
User Badges:

do you have the lines

ccm-manager config server x.x.x.x

ccm-manager config

in your running-config.

I have had similiar experiences and found by removing these lines, only adding them when making changes, that my problem was solved.

hauchinango Fri, 11/30/2007 - 08:01
User Badges:

I do have those lines in my conifgs...I will look at taking those out to see if it helps. Thanks much

allan.thomas Fri, 11/30/2007 - 12:11
User Badges:
  • Blue, 1500 points or more

Specifically bind mgcp control and media to an interface, preferably a loopback interface if possible.

I have come across problems similar to those that you mentioned, which subsequently caused MGCP to bind to a dialer interface for ISDN Backup. When routing reconverged, and the ISDN was torndown MGCP was still bound to the dialer interface which was no longer reachable, consequently packets were black-holed.

When the problem re-occurs verify what the source address is for MGCP, run 'show ip socket' and confirm that the address is reachable. Resetting the gateway will force a download of the config and will re-registered with the highest IP address.

Hope this helps.



This Discussion