Windows Server 2016 DFS-Replication backlog issue
I have 6 servers running Windows Server 2016 File Server DFS replication.
I'm having trouble with my DFS File server replication, the report generated from the health report and
dfsrdiag backlog /smem:$src /rmem:$dest /rgname:$GroupName /rfname:$foldername
Result:
SourceComputer DestinationComputer Enabled MinimumRDCFileSizeInKB RdcEnabled State BackLog
SERVER-B SERVER-C TRUE 64 TRUE Normal 226450
SERVER-B SERVER-X TRUE 64 TRUE Normal 243177
SERVER-B SERVER-N TRUE 64 TRUE Normal 227989
SERVER-B SERVER-A TRUE 64 TRUE Normal 227281
SERVER-N SERVER-X TRUE 64 TRUE Normal 322091
SERVER-N SERVER-B TRUE 64 TRUE Normal 321380
SERVER-N SERVER-A TRUE 64 TRUE Normal 2
SERVER-N SERVER-C TRUE 64 TRUE Normal 6
SERVER-C SERVER-B TRUE 64 TRUE Normal 625585
SERVER-C SERVER-A TRUE 64 TRUE Normal 255773
SERVER-C SERVER-N TRUE 64 TRUE Normal 298674
SERVER-C SERVER-X TRUE 64 TRUE Normal 626278
keeps on increasing, not decreasing. I have checked, that all of my File Servers are online and with more than 40% of free disk space capacity.
Any help to force the File server replication from one source to multiple target servers will be greatly appreciated.
Thank you in advance.
So as already mentioned, the backlog increase can be caused by staging quota being too small. To narrow down the issue, you should be able to trace this by 4202, 4204, 4206, 4208, 4212 events. These can be logged on source or destination servers. There are other possible reasons for the issues you are seeing. Here is an article on top 10 DFSR slow replication reasons: https://docs.microsoft.com/en-us/archive/blogs/askds/top-10-common-causes-of-slow-replication-with-dfsr
And some additional reading you might want to check on tuning DFSR replication speed: https://docs.microsoft.com/en-us/archive/blogs/askds/tuning-replication-performance-in-dfsr-especially-on-win2008-r2
I used to work with DFSR in one of the clients locations and seen complex issues with the replication. We were forced to use single namespace of DFS but the replication is what I wouldn’t like to use again. We decided to switch from the DFS replica to a lower level storage (block level) replication between the nodes and using shared storage for DFS. For the SDS stack, we used Starwind vSAN: https://www.starwindsoftware.com/starwind-virtual-san-free