DFS-N will not fail over

Two brand new Dell servers, Win 2008 R2, all patched. Both are DCs in new domain. Set up DFS-R, folders are replicating without issue. Both are at the same site and connected to the same switch, everything.

Set up DFS-N with both servers as nameservers. Namespace works fine - when users go to the namespace address, they see the files. Plus since the namespace is pointing to folders replicated with DFS-R, changes are replicated between the servers. So far, so good.

When I attempt to test by disconnecting one server from the network, users who were pointed to the "down" server remain pointed there (at least according to the client's DFS tab in folder properties), and I am unable to change settings on the "up" server (nothing can be changed while the other namespace server is down).

How to configure the namespace servers for high availability? Someone suggested clustering, but I thought the whole point of multiple namespace servers is to prevent this need for clustering.
milhouse537Asked:
Who is Participating?

[Webinar] Streamline your web hosting managementRegister Today

x
 
Joshua GrantomConnect With a Mentor Systems AdministratorCommented:
I believe that even though it is a namespace server, the rules still apply. A way to test would be to disconnect server 2 and see if server 1 hangs the same way.
0
 
Joshua GrantomSystems AdministratorCommented:
DFS and Namespaces are not meant to be used for High Availability, it is just meant for keeping files on multiple servers synchronized across slow connections.

Clustering is the way to go for High Availability
0
 
milhouse537Author Commented:
Clustering is not an option, unfortunately. What's the point of multiple namespace servers, then? Surely there must be a way to utilize this.
0
Problems using Powershell and Active Directory?

Managing Active Directory does not always have to be complicated.  If you are spending more time trying instead of doing, then it's time to look at something else. For nearly 20 years, AD admins around the world have used one tool for day-to-day AD management: Hyena. Discover why

 
Joshua GrantomSystems AdministratorCommented:
Having multiple namespace servers are usually used to split the referral load between multiple servers in very large environments. Usually a DFS-N server can handle thousands of referrals per second, but when you have a large environment and referral numbers get really high, its good to have a secondary to help process.

I would take a look at your DFS-N config, here is a good post explaining the utility.
http://blogs.technet.com/b/josebda/archive/2009/07/15/five-ways-to-check-your-dfs-namespaces-dfs-n-configuration-with-the-dfsdiag-exe-tool.aspx
0
 
milhouse537Author Commented:
So there is no failover for that at all? It seems like if one of the servers go down, it breaks the while thing.
0
 
Joshua GrantomSystems AdministratorCommented:
There is a kind of failover but not what you are wanting. Once the referral is cached to the client from a namespace server, and then that server goes down, it takes time for the failover process to start and for it to find another target server to service the referral. This time is dependent on what you set your root and link referral timeouts to.

Here is another article, it is older and says Server 2003 but the idea is the same

http://blogs.technet.com/b/filecab/archive/2006/01/20/417832.aspx
0
 
milhouse537Author Commented:
Thanks for the info, that is interesting. However the problem I'm having is that when server #1 goes down, it seems like server #2 locks up when looking at the namespace settings.

The moment I disconnect the network from server #1, I go into server #2 and I can't view any settings for the namespace or do anything else. When I try to go to target folder properties the MMC just kind of hangs and I get an error message eventually.

Are you implying, based on the link you mentioned, that server #2 will be subject to the same TTL as clients? I would think that as a namespace server it would be able to refer to itself.

Would waiting for the TTL to expire solve this issue with server #2? The failover doesn't have to be seamless. I'd just like the ability to keep operating somewhat while the problem is resolved is the downed server. And my testing hasn't been too successful so far :-(
0
 
milhouse537Author Commented:
Indeed, it does look like it goes both ways. I guess I'll try to decrease the TTL and see if the problem is resolved. Thanks.
0
All Courses

From novice to tech pro — start learning today.