UCCX 12.X Clustering Help

[aniket0422]

Hello Guys,

Is it possible to have 3 Nodes in a single UCCX Cluster, As per my understanding UCCX works only in HA mode in which we can have only 2 UCCX Servers.

Or else is it possible to configure cluster between two UCCX Clusters.

We have three Locations - A,B and C and contact center agents at all 3 locations.

CUCM node is there at each location. But UCCX is only at Location A and B as per HA design.

If location C gets isolated from network, there is no use of UCCX HA setup for the contact center agents of C location.

Is there any way we can mitigate this scenario.

[Chris Deren]

CCX Cluster can only have 2 nodes, they are setup in Active/Standby fashion. This does not work like CUCM where all nodes are active at a time, thus all calls/Finesse connections/etc will connect to the active node.  You can split the 2 nodes across WAN if you meet the design requirements outlined in the CCX SRND.

[Anthony Holloway]

Technically you could login Finesse to either UCCX node now. That was a recent change in 11.6.

https://www.cisco.com/c/en/us/td/docs/voice_ip_comm/cust_contact/contact_center/crs/express_11_6/release/guide/uccx_b_release-notes-for-uccx-116/uccx_b_release-notes-for-uccx-116_chapter_01.html#topic_6EF9FDD9B1BFBE4760334AE53352C708

[Anthony Holloway]

While you can only have 2 UCCX servers in a cluster for redundancy, you can have multiple UCCX clusters tied to your single CUCM cluster for different reasons. Such as: capacity, fault isolation, version control, and geographic locations.

[aniket0422]

So I can have two UCCX Clusters linked with Single CUCM Cluster.

What will happen if User from UCCX Cluster A moves to UCCX Cluster B. What challanges we may face.

[Chris Deren]

Since both CCX clusters are independent of each other, you would need to provision the agent separately on both clusters and then have agent login to separate finesse URL.  Agent's phone would need to be associated with both RmCm application users (one for each CCX instance).  Any call flows, reporting would be tied to single cluster, so you would not be able to queue concurrently across agents on both system, you could create some kind of overflow redirect by transferring calls between systems if needed.