Currently in HVR you have an option where if drop table is seen HVR only drops it internally but not on the target. 

Unfortunately we have seen in a number of scenarios where drop and recreate has been detected by HVR which isn't genuine, as such a drop & refresh is triggered, which then leads to multiple hours and in some cases multiple day outages in accessing those tables.  

As such we would request a feature where if HVR detects on the target it needs to actually drop and recreate the table and then refresh it, the integrate goes into an error state and the drop is not triggered (possibly print an error that integrate cannot proceed without a drop and recreate or with running x cmd line to skip it ). 

This will ensure all encounters of an unexpected drop and recreate are not actually fired. Yes we can do this via failed integ job on EnrollBreak. However we would still want all other adapt DDL changes to be applied i.e column additions etc. We just only want to stall out drop and recreates. 

Our Sources include Sybase & MSSQL  and our targets include Sybase MSSQL & Snowflake