If you use user-generated content, such as content sets that contain probabilistic or classifier models, verify that the Content Management Service on the backup node runs continually. When the Content Management Service on the backup node runs continually, any change to the user-generated content on the primary node is copied to the backup node.

The Content Management Service properties on the backup node can differ from the properties on the primary node. For example, the address reference data and identity population data locations can differ.

When the Content Management Service on the primary node is down, you cannot import reference table data from a flat file or a relational table, and you cannot create a probabilistic or classifier model from a flat file. To import flat-file data, restart the Content Management Service on the primary node. Or, reconfigure the master status of the Content Management Services on the primary and backup nodes, so that the Content Management Service on the backup node becomes the master Content Management Service.