We are configuring HVR in one of SQL DB server , here we have some queries ,
1. publisher and distributor can be delete right after creating the article. Why we need distributor database for HVR?
2. if we failover database from PR( datacenter) to DR ( datacenter) this existing architecture will not work
1. The distributor database is a core part of SQL Server Replication, and so is the concept of articles. We need articles to enable what we call supplemental logging so that on the destination side the SQL can be replayed using primary key columns. Given the two concepts are tied together I would strongly recommend against deleting the distributor database after articles were created.
Note that HVR 5 by default uses CDC tables to track changes and if you only use CDC tables to track changes then a distributor database is not required and does not need to be created.
2. If you refer to the distributor database then please note that in the context of HVR this database is simply a side-effect of the use of articles and there is nothing in the database that determines the routing of the data. HVR takes care of this.