Multiple datastores can be aggregated to create a datastore cluster. A datastore cluster is a collection of the member datastore resources with a shared management interface. vSphere Storage DRS manages the datastore cluster resources to determine the initial placement and ongoing balancing of virtual machine VMDKs across the datastores in the cluster. Datastore clusters are supported for both VMFS and NFS datastores.

A few recommended practices when using datastore clusters and Storage DRS are as follows:

• Cluster datastores with similar IOs and capacity characteristics
• Use separate datastore clusters for replicated and nonreplicated datastores
• Do not mix NFS and VMFS datastores in the same datastore cluster
• Do not place datastores shared across multiple data centers in a datastore cluster

When a virtual machine is placed on a datastore cluster, Storage DRS determines which datastore in the cluster the files will be stored in, based on space utilization and/or performance. The following diagram is a logical representation of the virtual machines
placed on a datastore cluster:

The best practices when using datastore clusters are as follows:

• Datastores in a cluster should have similar performance capabilities
• Keep similar virtual machine IO workloads together on the same cluster
• Do not mix replicated and nonreplicated datastores in the same cluster
• Do not mix NFS and VMFS datastores in the same datastore cluster
• Use VMDK affinity rules to keep virtual machine disk files together on the same datastore within the datastore cluster
• Use VM anti-affinity rules to ensure that virtual machines run on different datastores within the datastore cluster

The VMware vSphere Storage DRS Interoperability whitepaper can be found at http://www.vmware.com/files/pdf/techpaper/vsphere-storage-drs-interoperability.pdf. This whitepaper provides an overview of the datastore cluster best practices and interoperability of datastore clusters, along with other VMware products.