We're designing an infrastructure that calls for several clustered pairs of mailbox servers. However, because of a slipup on the part of our purchasing department, we'll have only half the servers by the deployment deadline. Can we install those servers as single-node clusters, then upgrade the clusters when we get the remaining servers?
A single-node cluster might seem as oxymoronic as "jumbo shrimp," but the configuration actually makes sense in some cases. Your situation is a perfect example—you need to set up the cluster now but want to be able to add nodes later. So, yes, you can install the servers you have as single-node clusters, and when your purchasing team gets the additional servers, you can simply add them to the clusters without interrupting normal operations. In the meantime, you won't be able to take advantage of many of the benefits that a cluster's failover capability offers, but you will be able to easily restart services (you can just restart the virtual server) and host multiple Exchange instances on one server.
