I have a three node cluster (Win2K8R2) with 3 instances of SQL Server 2008 (each on one node of the cluster.) Two of these instances need to participate in MSDTC communication (I believe that one sends to the other.) In reading the agent guide it mentions:
MSDTC servers can co-exist with SQL servers on the same cluster nodes. If the MSDTC Server and the SQL Server are running on the same node, the MSDTC client is configured in the default configuration. If the MSDTC Server is not configured on the same node as the SQL Server, then the MSDTC client must be configured on that node. In general, you must configure the MSDTC client on all nodes except the node on which the MSDTC Server is configured to fail over.
I don't know how this applies to me as I'm not really that familiar with MSDTC. Do I need an MSDTC server in a separate SG (and thus be limited to running it on a node that doesn't have a MSDTC client configured?) - and I think in that scenario I need another cluster node?. Can someone help sort this out?