I have DFS setup on my network as follows:
- A domain we'll call DOMAIN.LAN.
- Two AD controllers, let's call them DC01 and DC02.
- Two file servers, let's call them FS01 and FS02.
- A Windows 10 Pro workstation joined to the domain.
On the file servers, I have a share called "public", which is successfully accessible from the Windows 10 workstation as \\FS01\public and \\FS02\public.
On the AD controllers, I have a DFS namespace set up called "dfs", with a share called "public". This share has the following targets:
- \\FS01.DOMAIN.LAN\public
- \\FS02.DOMAIN.LAN\public
I've tried accessing the share from the Windows 10 workstation in the following ways:
- \\FS01\public: This works 100% of the time.
- \\FS02\public: This works 100% of the time.
- \\DC01.DOMAIN.LAN\dfs\public: This works 100% of the time.
- \\DC02.DOMAIN.LAN\dfs\public: This works 100% of the time.
- \\DOMAIN.LAN\dfs\public: This works intermittently.
It's this last one that's the issue. When I say "intermittently", I don't mean that I click on it once and it works, then click on it again and it might not. I mean that it may work for several hours, then stop working for several hours, then with seemingly no changes start working again.
I've tried changing the targets from the FQDN of FS01.DOMAIN.LAN to just FS01, or even just the IP address of the file server, and this makes no difference. However, since I can access the share through the DFS name space if I hit the AD controller directly rather than the domain name, this would seem to rule out both DFS configuration problems as well as DNS problems (even though the fact that it's intermittent would also seem to point toward DNS), which is why I'm unsure on how to proceed with troubleshooting.
A couple of things to note that may or may not be relevant:
- DC01 is running Windows Server 2019.
- DC02 is running Samba 4.9.5.
- FS01 and FS02 are running Samba 4.9.5 and are domain member servers.
- The Windows 10 workstation is connected to the network with the servers over a VPN. However, the IPs of DC01 and DC02 are (for testing purposes) hard-coded into the workstation to make sure that it can resolve DNS properly. Pinging all of the above FQDNs works as expected 100% of the time.
I do recognize that the VPN could be an issue. However, a workstation that is also local to the servers exhibits the same symptoms, so I don't see the VPN as a differentiator. The VPN connection to the server is solid as are both internet connections, and ping times across the VPN are 20ms or less.
I'm looking for either someone that has experienced this issue and has the solution, or at least a process I can use to further troubleshoot.