Question : How to Configure Failover Clusters With Win 2008 Server R2?

Dear Sir,

Windows Server 2008 provides two clustering technologies: failover clusters and Network Load Balancing (NLB). Failover clusters primarily provide high availability; Network Load Balancing provides scalability and at the same time helps increase availability of Web-based services.

My Scenario is the following:

I have 6 Terminal Servers (RDS) Win 2008 R2 X64, which I have an application installed on those Servers, Students connects through Thin Clients to those server for a session, then the application which is installed on the RDS servers launch automatically, and connect to SQL database Server...
DC is also available for authentication.
SQL server is configured mirrored to another SQL server for DB high availability...

Which solution works better for my scenario, Failover Clusters or NLB for the RDS servers (TS Session)?

In the past I used to configure (NLB) on the TS servers with Win 2003 server to load balance and failover...

Now, I moved to Win 2008 Server R2 (RDS), I was wondering if it's possible to install and configure Failover Clusters instead of NLB to provide high availability of my RDS servers incase if one server goes down...
The application is an exam app for studetns which should be available all the time during an exam...

How can I configure Failover Clusters for the 6 RDS Servers, I need a step by step guide or screen shots on how to do this?

Thank You!

Answer : How to Configure Failover Clusters With Win 2008 Server R2?

Sorry I am not explaining myself to well.  I'll try better.

Clients do not depend upon the session broker function to be able to connect to the terminal server farm.  In your case, when using a NLB cluster the clients will be connecting to the virtual IP assigned to the cluster.  As they connect, the session broker acts as a 'man in the middle' to direct the client to a particular server, or reconnect to an existing session on a particular server.  Now, say for example for whatever reason the session broker is down; clients will still be able to connect to the NLB cluster, but not have the session broker functions (i.e. direct to an existing session).

Placing your session broker role service on your DC should be ok, I was just trying to stress the point that if the physical box went down then you would have bigger concerns with your network other than the session broker service not working.

The session broker does indeed support connection to existing session, thats basically the purpose of it.  However of the 3 examples you give the only one that would work there is thin client rebooting.  Don't think of this as clustering, its not.  The session broker basically sits there and as clients connect it does a check to see if there is already an existing session open on one of the servers, if there is then it reconnects the incoming user to that server, if not then it puts them on the next available one.  If your terminal server goes offline, then all the sessions on that server will die with it.  As the users reconnect, they will be treated as a new incoming session.  The session broker will know that the terminal server is down and will not direct users to that server until it is available again.

Now, I believe you need to have 1 08R2 server as a licensing server, however I am not 100% certain on this.  I haven't tried to configure GPO's for 08R2 servers, I would imagine it should still work, but since you need an R2 server as your licensing server you should do your GPO stuff on this server.

Client side connects to the NLB IP, yes.

DNS round robin is only meant if you are not using NLB.

You still need to configure the farm name on the terminal servers, but this should be done from within the GPO you create.

Follow the steps outlined in the technet article I linked to above, they will be much better than anything I'll be able to muster.
Random Solutions  
 
programming4us programming4us