Add DFSR Member to replicate with 2 existing members

Copper Contributor

Current Situation. Replacing some servers and adding some for DFSR/Namespace in our environment. 

2 old servers 2012 R2

Splitting those off into 3-4 more 2016 servers. 

Server 1 replicates with Server 2 and vice versa (2 way replication)

Most of the time people connect to Server 1. If they VPN in they will hit Server 2. Data between the 2 servers seems off a little.  Not sure the best way to check. When I attempt to run diag or health check for DFS it times out or never finishes. So that is unreliable right now. 


Currently added new server to replicate with Server 1 and I have most of that done. Adding another server to replicate with Server 2 which I will do. 


In the end I need to have Server 3 and Server 4 which now have respective connections/replication to their older counterparts Server 1 is replicating to Server 3 and Server 2 is replicating with Server 4. Nobody is touching Server 3 or Server 4 until we are done with this project. 


How can I safely have Server 3 replicate with Server 4 not knowing for sure if data on Server 1 and Server 2 is the same? I am fearful if I do that, that data will be removed on Server 1 or 2 if for some reason the data is not on one or the other when syncing for replication. 


Does anyone know if I can trust doing this or is there something I should do to check and be sure? Basically if I add servers into the mix as a mesh connection I need to make sure all servers have the same data or fear of losing data in the process. 

Thank you to anyone reading this!

2 Replies



Check the POST marked "Answer" on this one. It has a lot of tools, etc. that might help you at least check. As far as the other information, I need to read it again to make sure I understand it before attempting to answer. :)


@Edward Dake  Hi Edward. Sorry for the delay in my reply :-). I checked out those tools and they will help yes. I have backlog monitoring/scripts already so I check that frequently. Health reports don't seem to generate however as it takes forever and times out. I guess that option isn't working now so maybe once I finally cut over to newer boxes it will work hehe.


Scenario again is Server A and Server B are syncing together. I created server C  and D which both are now replicating with C to A , A to C and D to B, B to D respectively. I now need to basically sync, D to C. One if these is less used as far as what server people are connecting to when updating files etc. 


Would it be safe to assume and trust that if I added a new member to a replication group that's currently replicating A to B, B to A , C to A and A to C now add in D to C and C to D? I would think since both servers have the data that they sync if any would be very minimal and wouldn't require to much back and forth. But that is my assumption.  What are your thoughts on that approach?