MGCP VGW Route List failover failure

Unanswered Question
Jan 27th, 2010
User Badges:

I am working with a CUCM system that has approx. 30 end nodes.  Each node has a VGW using MGCP and a PRI.  I have a route group configured for each node with there PRI in the group.  Then I have a route list that has each of the route groups in the list.  Here's the problem.  When a broadcast message (via IPCelerate's IPSession i.dialout) is initiated, the first call goes out of the first route group entry in the route list, as expected.  Subsequent calls go out as expected as well, until the first route group PRI is full.  Once we get to the point where there are no available ports on the first route group, I would expect that the second route group is engaged to continue making calls.  This is not the behavior I have experience.  Instead, I receive errors shown in the call log on IPSession with cause code (101, unknown).  I have researched this error and it turns out this error means that there were not available channels to make the call.


Question:  Once the first route group PRI fills, why does CUCM not fail over to the next route group to continue making calls?


From my current research, I find documentation that multiple PRI entries in a single route group will accomplish this, but nothing that deals with failover on the route lists themselves.  Our original understanding of the route planning was to use the group to specify a node and then lists to create failover scenarios.  The above does not act this way.  Do I need to reconfigure my route plan to redefine failover in the route groups?


Thanks, in advance, for any assistance.


Dom Rozzi

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
markbatts Tue, 02/02/2010 - 02:34
User Badges:
  • Bronze, 100 points or more

Hi,

Just so i'm clear, you have a Route list with multiple route groups in.When the Gateway in the first route group is busy the calls should then be presented to the second Route group.If this is correct then what you are trying to do is very normal standard configuration which people use and should work.What version of callmanager are you using?


mark

dom_rozzi Tue, 02/02/2010 - 04:12
User Badges:

Mark,


Your assessment of our design is correct. We just moved to 7.1.3bsu1, but

our problem was discovered in 6.1.4. For some reason when we push 100 or

so calls out at one time, the first 15-20 get out and then the a series of

calls fail and then another 15-20 go out and so on.

Thanks Dom

markbatts Tue, 02/02/2010 - 04:53
User Badges:
  • Bronze, 100 points or more

It sounds like that its something to with the amount of simultaneous calls you are placing not the configuration.You may need to speak to cisco regarding this if you have the same issue in 6.14 and 7.1.3.Sorry can't help anymore on this one

Actions

This Discussion