We have a SQL Server 2008 database that is replicated to Australia. The primary data tables are merge replicated. The replication cycle is 1 hour. We have a situation where a user inserted a row into a table, deleted the row, then reinserted it within an hour. The merge replication failed because of a primary key uniqueness violation on that table. We have been told that merge replication cannot handle this scenario.
From what little I can get from the MS literature, the replication triggers should have handled this. The delete trigger should have removed the insert entry from the replication tables and the second insert should have removed the entry from the delete replication tables, leaving the replication processing to only do the second insert. Can anyone clarify what is really supposed to happen here or if this is in fact a limitation of merge replication?