Showing results for 
Search instead for 
Did you mean: 

Unity Connection 8.5 Cluster Call Distribution Issue


I have Unity connection 8.5 Cluster.

I am trying to test the redundancy and failover by shutting down the Publisher server.

The subscriber successfully takes over the Primary status.

The Problem is that once i shutdown the Publisher server and the subscriber takes over the Primary role, i stop receiving calls on the Auto Attendant which means that the calls are not handled by the subscriber server. I have 96 unity ports and i have registered half with the Publisher server and the other half with the Subscriber server.

The question is how to make the Subscriber handle calls once the publisher server goes down?

8 Replies 8

First off the SRND for Unity Connection recommends that the subscriber take calls first and the publisher back it up. While both are in service the publisher is intended to perform integrations (e.g. Single Inbox to Exchange) and serve client requests (e.g. IMAP, RSS, VMO, CiscoPCA).

If you monitor the call from RTMT or the rPSM do you see the call arrive on CXN at all? If not then you want to look at the CUCM side:

  • Did you assign half of the ports to the subscriber?
  • Are all the ports registered? Does CUCM show the CXN subscriber IP address as the client registered IP address for those ports?
  • Are the ports of the subscriber in the line group for CUCM to route calls to?

  • Did you assign half of the ports to the subscriber? --> YES
  • Are all the ports registered? --> YES
  • Does CUCM show the CXN subscriber IP address as the client registered IP address for those ports? --> YES
  • Are the ports of the subscriber in the line group for CUCM to route calls to? --> YES

I actually what i am trying to achive is to have the cluster handle the incoming calls, to be distributed between Publisher and Subscriber, or to have the subscriber take all the incoming calls but in this case what will happen if the subscriber goes down?

Dear Jonathan,

any update on the above as i have answered your questions.

Thank you and waiting for your response.

The failover of Unity Connection is handled automatically by it's active-active failover model.  The routing of calls is more of a CUCM configuration.  I split the ports into 2 line groups.  Ex:  Pub_LG has all of the Publisher ports assigned.  Sub_LG has all of the Subscriber ports assigned.  (Note: only ports that are configured to Answer calls in Unity Connection should be assigned to their respective line group).  LG's are configured as Top Down with failover to next member then next group.  From there, you create a Hunt List which has both line groups with the Subscriber being preferred.  Ex:  VM_HL with Sub_LG, Pub_LG.  Then you just need a single Hunt Pilot which points to the VM_HL.

Can you elaborate on how the CUCM configured in your environment to distribute calls?  The recommendation is that, for a cluster, the Subscriber be configured to take the majority of calls during normal operations (in other words, you are routing calls to it first in CUCM with failover to the Publisher as noted above).  If you have this sort of configuration already and failover is not working as you would expect then it may be worth looking further into the actual details on the Unity Connection configuration - such as specific version you are running, etc.


Please rate helpful posts!

I have the same issue and tac is telling me the same thing. Create separate line groups for each unity connection server and create separate port groups in unity connection. The problem is I cant find this configuration documented anywhere. The doc Ive read states to just create one line group with all of the ports and one port group in unity.
What your saying makes perfect sense Im just wondering why its not documented anywhere
Sent from Cisco Technical Support iPad App


On the documentation part, I can't answer that question...personally, I've always done it that way.  There is a SCCP integration guide, which may (or may not) refer this particular configuration...have you reviewed that document?



The SCCP integration guide is the guide Ive been referring to. I didnt see anywhere in there about using 2 lines groups one for each server. I resolved our issue. The issue for us was with the line group hunt options being mis-configured. The "Not Available" opition was set to stop hunting so when the first port is attempted, which is on the Sub, is not available the line group stops hunting. So for us when the sub is down no ports were working.

Thanks for all you and others like you do on these forums. Im often times able to find answers from you guys a lot faster than digging through the documentation.



The way I had this issue solved is by dividing the port license for both servers.

which means that i had 1 licenses installed on the Publisher server for the ports and the TAC told me to split the license file into 2. 1 file will be installed on the Publisher and 1 on the Subscriber.

By doing so the issue was solved.

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Recognize Your Peers