Administration is carried out on the server, using the graphical tool? Server Manager? Windows Server 2016 operating system or through the command line. During the installation of the Storage Space Direct Cluster, first of all, all the hard drives of the different servers are grouped in a Storage Pool . It is important that only those disk drives that contain user data are integrated into the cluster. The server operating system disk drives remain independent.
Suitable hard drives and semiconductor storage media are automatically detected and added. In this way, later, more hard drives and servers can be integrated into the cluster to scale the storage capacity as needed . Each node supports four data disk drives and two cache disk drives in addition to the system hard disk.
Virtual disks are created in the new cluster and are here called Cluster Shared Volume (CSV). These CSVs are then formatted, a process that allows for two file systems. Microsoft recommends CSV-ReFS (Resilient File System), because it is better suited to Storage Spaces Direct technology than CSV-NTFS, based on the classic NTFS for hard drives. CSVs can span multiple hardware disk drives, but this does not affect the user since they are all managed by the system.
Each Storage Space Direct node can reach a total storage capacity of up to 100 TB. A multi-server Storage Pool can only use a maximum capacity of 1 PB (PetaByte = 1024 TB), even though Storage Spaces Direct (S2D) allows up to 16 servers.
The network communication between the different nodes is done through the SMB3 protocol , an evolution of the Server Message Block (SMB) with SMB Direct and SMB Multichannel.