We have an implementation that we are testing of 5 Windows 2016 RDS servers utilizing FSLogix for profile containers. The User Profile Containers for this RDS farm are located on 2 Windows 2016 File Servers that utilize a DFS-N and DFS-R, so user Profile Containers are replicated across both File servers.
My question is what are the best practices for rebooting the backend File Servers that house the Profile Containers? I'm experiencing issues when users are logged on to the RDS servers (either with disconnected sessions or if they are still working) they lose connection to their Profile Container disk when the Backend servers are rebooted. I would have thought that DFS would kick in and reconnect the user with their Profile Disk on the 2nd File Server but this is not the case. The only thing we are able to do is log the users off once the backend server has been rebooted or reboot the RDS servers.
As we need to reboot the backend servers for patching and maintenance is there a best practise for dealing with reboots to the backend storage of FSLogix Containers?