Inconsistent replicated data across replication nodes

There are various ways to detect that replicated data is inconsistent across nodes. For example, you can use ad hoc queries that return different results, row counts from replicated tables, or the qsum analysis package, which is included with the NPS®.

If replicated databases are inconsistent across the cluster, determine which version of the data to keep. Typically, the content of the primary replication node is chosen. To recover the set of inconsistent databases on all applicable nodes, use the replication recovery tools.