Data Archiving [Design]

I am developing an archive module for an application using Dotnet and SQL Server as back end. From multiple approaches of archiving we've decided to build a custom application to archive the complete database up-to a chosen threshold to another mirrored database and then removing the archived items from source DB. This has to be done from a custom application and not from a Linked Server, job, SSIS, Replication or any thing else.

There are so开发者_如何学运维me items which will be emptied and rewritten again on each archival run. Before starting to build the module we were thinking to remove the Foreign Key constraints from the mirrored archived database to avoid any referential integrity violation when we Emptied these tables and re-write (Only constraints will be removed, columns and values still exists in the archived DB). However this approach not even seems fishy to me but kinda disturbing as well. So here I am asking what will be the right approach to deal with this?

We stick to this design in the question.