DFS Replication has stopped working

[JDkfc64]

This is the warning:

The DFS Replication version vector size has exceeded acceptable limits. A large version vector size could cause degraded DFS replication performance, poor responsiveness of DFS replication management operations, and excessive memory and CPU resource consumption. Contact Customer Support Services to analyze the overall health of your DFS Replication deployment.

Have tried recreating the replication group and deleting the DFSR folder in the system Volume information folder on the drive

And increasing the staging quota for the replication group

[Jay-Updegrove]

First up, welcome to the community!

Second, please give us some details. What OS is DC running, what’s the forest level, how many DC’s do you have (virtual? physical? both?)

The more detail you can supply here, the better our chances of helping!

[Nerf_Herder]

Also include with the OS the build version and did this happen after a recent Windows update? IF so, can you back it out?

[JDkfc64]

Thanks Folks. Forest Level is win 2012R2. The OS on both source srever and destination server is 2022 standard version: 21H2, Build 20348.3692.
There are 13 DCs with various OS from 2012 R2 to 2019 standard, 4 of the DCS are virtual.

[JDkfc64]

The destination server is a dedicated DFS server, so data only being replicated one way to the destination server . DFS is working on this server for other source servers (replication groups), but just one or two replication groups not working. No stand out error in logs

[Jay-Updegrove]

Wow…13 domain controllers? I’m guessing this is a massive company? Can you isolate where the failure is coming from? Any one of those 13 giving you any indication of problems?

Is it possible to do away with any of them? It’s fine to have a single, Read-Only DC as a VM at each location unless you just have to do a full DC, and cut out a lot of the complexity.

You’ll want to try and isolate what’s different between the working and not working groups. Solve for X wherever possible. With replication failures, it’s almost always one minor detail that’s different on DC 1 vs 2. And again, try to simplify your structure when and wherever you can. If you have the resources at the sites where DFS is failing, stand up a new VM, make it a RODC (if applicable), demote the ‘failing’ unit(s) and see if it helps having everything at the same OS level.