Listener failover

This topic pertains to a clustered listener configuration. As long as at least one Unica Campaign listener is viable, failover should occur without disruption.

Failover involves automatically switching to an alternate node in a cluster. Listener failover may occur for any of the following reasons:
  • Network issues (TCP/IP)
  • Listener (software) failure
  • Hardware failure
Failover ensures that when a listener node is unresponsive for any reason, another node in the cluster takes over. Whenever possible, any flowchart sessions (unica_acsvr) that were spawned by the failed listener are also recovered, so flowchart work is not lost.

In rare cases, there may be a non-recoverable situation and any work that is in memory will be lost. In this case, a message alerts users to the situation so they can make a note of any flowchart changes that need to be redone.

To avoid possible loss to flowchart work, best practice is to configure checkpointFrequency and autosaveFrequency in the Unica Campaign partition settings. You can override the global configuration settings for individual flowcharts, using the Admin > Advanced option.