I have Azure SQL Georeplication setup on a production database. This process has worked fine for some time. The georeplicated database is on another server in an Elastic Pool. I now want to keep an on-prem database in sync and am considering setting up Azure SQL Data Sync (ADS). My preference would be to have Azure Data Sync work between the georeplicated replica and the on-prem SQL. In that model, Azure Georeplication keeps working as is, and the data from the relica syncs to the on-prem server.
I have read all of the limitations of Azure Data Sync and it does not say that this model will not work. In my case, I have over 500 tables in the data model, only 25 or so that need to be synced. When I use the Azure Portal to setup these tables, it errors saying that ADS only supports 500 tables. The official workaround for this is to create a user that only has rights to the tables I want to replicate (the 25), and use this user to connect to the hub DB.
Because the replica is read only (since it is a mirrored reeplica), the CREATE USER statement fails, complaining the DB is read only.
My question: Is this a feasible scenario, to use a georeplicated database as a hub DB in Azure SQL Data Sync? I can create the user in the base DB which will replicate to the georeplicated DB, but it feels like plumbing, and worse, when I add my replicated tables (the 25) it will need to create triggers on those tables which will also likely fail because of the readonly attribute of the replicated DB.
Any insight is appreciated.
Thanks...