High Availability (HA) / Duplex Deployment

Overview

Microsoft Dynamics Web Connector supports high availability(HA) for Finesse provided that Cisco Finesse is configured to support HA. If Finesse supports HA, it will be hosted on at least two servers i.e. primary and secondary.

At any given time, one of these nodes will be acting as the master node and the other as the slave node. If both nodes were up and for some reason the master node goes down, the connector will automatically connect with Side B, and an agent will be able to continue his call, however, there are certain limitations which are mentioned below

Note

Once the active finesse is down, the connector tries 3 times to connect to the same server, after that, it will show the disconnection error. Once the connection is disconnected, it then checks for an available server and connects with it. The whole process takes roughly 3 - 4 minutes.

UCCE Issues/Limitations:

• The agent state after failover becomes Not Ready - Connection failure. This behavior in different agent states prior to failover is described in the following table.

UCCX Issues/Limitations:

The agent is on an active Call and after that agent performs the following action on the same call

• Call Type internal for conference case as not able to track the origin of the call.
• If the Agent state is Ready before failover, the agent state becomes Not Ready - Agent Logon after failover.
• Two call activities will be created for Agent A when direct transfer on Agent B after the connection is restored.
• Consult case popup on the agent's number instead of the customer's. For B always, it could also happen for A if we receive a consulting event first.
• The call variable null issue affects activity for a record associated with the activity.
• Direct transfer issue, null in connected number and no pop-up. (We get the number from the call variable in this case).

Finesse Limitations:

• Agent State changed to Not ready (Agent Logon) after the call ended / during the call.
• The supervisor is unable to silent monitor/barge in after the connection is restored.
• Brage in button gets disappears on Finesse after the connection is restored (when failover during silent monitoring).
• Call variables become null.
• If we login the agent on slave node and then failover happens, it doesn’t switch to other node.

All issues are related to Call Data not being available after failover.
Cisco Documentation Link: Cisco Unified Contact Center Express Administration and Operations Guide, Release 11.6(1) - Unified CCX Introduction [Cisco Unified Contact Center Express] - Cisco

"If the active server fails, the Unified CCX provides automatic failover to the standby server. If the active server fails (for example, in the event a hardware failure occurs or the Unified CCX Engine process terminates), some calls being handled by the server are lost. The lost calls are restricted to those being handled by the system (those in the IVR stage or in the queue). Calls answered by agents continue to remain live even though related data on the agent desktop is lost. "

Note

The above issues/limitations are only for the current ongoing call on which failover occurs. For the next call, all the operations should be performed smoothly.