Showing results for 
Search instead for 
Did you mean: 
Frequent Contributor

Question on Extension Mobility in HA deployment

We're running a CUCM 8.6 cluster and we've configured about 150 users with Extension Mobility.  As part of the server configuration, you need to configure URLs in System > Enterprise Parameters and a URL for the EM service itself.  We currently have these configured with an IP address to avoid a system dependency on DNS (per best practices on deployment), however, this presents a challenge when operating during a cluster node failover.

As is, the enterprise paramter URLs and the EM service URL are pointed at the CUCM Subscriber.  If the subscriber fails, this means that all IP Phone service URLs and Extension Mobility will be inoperable.

We're looking for a recommendation on mitigating this single point of failure, as we need Extension Mobility to be fully HA in the environment.  Should we bite the bullet and configure these URL's with a DNS-dependent hostname with a short TTL (so we can change the IP in an emergency), or is there a better solution here?  It's my understanding that Enterprise Parameters are cluster-wide, so even if we were to install a second instance of the Extension Mobility service on the IP phones with a URL to the other cluster node, we'd also have to configure multiple softkeys to them in order to avoid the dependency on the Services button.

VIP Mentor

Question on Extension Mobility in HA deployment

Hi Paul,

Cisco best practice for EM redundancy is to use SLB (server load balancing)

In order to provide redundancy for the Cisco CallManager Cisco IP Phone Services (or URL Services) and the EM IP phone service components, Cisco recommends using the Server Load Balancing (SLB) feature available in Cisco IOS to front-end multiple Unified CM nodes. The Cisco IOS SLB feature provides a virtual-IP-to-real-IP address mapping, that front-ends the real IP addresses of the Unified CM nodes. The SLB feature can be configured to monitor the status of multiple nodes, automatically redirecting requests during failure events. By using the SLB virtual IP address (or DNS hostname) for the URL Services and EM IP phone service, you can ensure that both components are still available during any node failure and, therefore, EM login and logouts will continue

Two Virtual IP addresses will be required. The first VIP will point to the Cisco CallManager Cisco IP Phone Services (or URL Services) and the second will point to the EM IP phone service. The VIP will be configured to resolve to the cucm real IPs in your preferred order e.g

  • •1.   10.x.x.x Publisher
  • •2.   10.x.y.y  Subscriber

Some people have used DNS to do this, but according to the statement below it is not recommended.

NB: According to Cisco SRND “Cisco does not recommend a redundancy design using DNS records with multiple IP listings. With multiple IP addresses returned to a DNS request, the phones must wait for a timeout period before trying the next IP address in the list, and in most cases this results in unacceptable delays to the end user. In addition, this can result in more than two subscriber nodes with the Cisco EM Application service enabled to handle login/logout requests, which is not supported.”

Please rate useful posts

"For the love of God is broader than the measure of man's mind And the heart of the Eternal is most wonderfully kind"

Please rate all useful posts
Frequent Contributor

Question on Extension Mobility in HA deployment

Thank you for the detailed (and super quick) answer.  I'll be looking into the SLB option.  It does technically introduce another single point of failure ot the environment, however, a router failure is far less apt to happen than anything running under UCS, as I've found in our environment at least.

In any cae, I can test it with a couple manually configured phones first, before having to change an enterprise-level setting.

Thanks again.

CreatePlease to create content
Content for Community-Ad
August's Community Spotlight Awards