VMware has a good whitepaper on setting up a Microsoft fail-over cluster in vSphere. Unfortunately there are number of restrictions when using Microsoft clustering with VMware vSphere, which really put a damper on the practices in certain environments.
The following environments and functionality are not supported for MSCS setups with this release of vSphere:
– Clustering on iSCSI or NFS disks.
– Mixed environments, such as configurations where one cluster node is running a different version of ESX/ESXi than another cluster node.
– Clustered virtual machines as part of VMware clusters (DRS or HA).
– Use of MSCS in conjunction with VMware Fault Tolerance.
– Migration with VMotion of clustered virtual machines.
– N-Port ID Virtualization (NPIV)
– With native multipathing (NMP), clustering is not supported when the path policy is set to round robin.
– You must use hardware version 7 with ESX/ESXi 4.0.
The major limitations which really stick out are the lack of VMotion, and a cluster which has DRS or HA enabled. Depending on your storage array, the round-robin multipathing limitation could be a real issue as well. I’m not sure whether these limitations are imposed by VMware because of technical glitches, or if Microsoft has imposed them. Either way, you need to be aware of the limits when you look at using MSCS, and they may drive you to other HA solutions.
Remember there are two aspects to high availability: the service and your data. MSCS does nothing for data high availability. For that you need to look at database mirroring. For service HA you have other options, such as vSphere HA and FT. But vSphere HA and FT don’t increase availability when you need to reboot the SQL host when doing planned maintenance such as applying patches or service packs.