Sometimes for availability purposes, it is necessary to ensure the MDT database is clustered to allow failover in case of a server failure or maintenance event. There are a few things that you want to keep in mind when you decide to do this:
1. The SQL Browser service is NOT cluster-aware. This means if you put the database on a named instance of SQL, when clients request a connection to the SQL Cluster IP, the node IP will respond. This makes the Windows Firewall in Vista, Win7 and even Windows PE regard the traffic as unsolicited and hence blocks it. You can add an exception to all of your Full OS clients, however WinPE does not allow you to customize the firewall and you would have to shut the entire firewall off – this is not recommended. The best way to resolve this is to ensure that the MDT Database is on the DEFAULT INSTANCE and does not need to use the SQL Browser service.
2. MDT Authentication share. In order to check windows authentication to the SQL Server, MDT attempts to connect to a Windows Share on the box. In a cluster, you can set this share up as a cluster resource. Ensure that the cluster resource fails over with the MDT database, i.e. is part of the same group. Also make sure that the network access account has access to the share and DB Reader access to the database. If you are not using windows authentication, ensure that you include DBID and DBPW fields in your queries.
3. Use the Cluster name in CustomSettings.ini. In the CustomSettings.ini file, ensure that the name you use for the SQLServer property in each query is the cluster name – do not specify an Instance Name.
4. Named Pipes – Ensure that you have named pipes set up on your SQL Server.
Note that this post covers clustering the MDT database only. You can follow LTI – Following Daniel’s excellent advice on clustering the resource of the Distribution Share http://blogs.technet.com/deploymentguys/archive/2008/03/31/up-ping-the-uptime-of-your-lite-touch-server.aspx you can ensure that this resource is also on the same group as the MDT Database.
Happy clustering :).
This post was contributed by Aly Shivji a consultant with Microsoft Services – U.S. East Region.