The VxRail P-Series nodes support up to 24 disks per node. Only all-flash configurations are supported for SAP HANA for the dual-socket P570F[1], while only NVMe configurations are supported for the quad-socket P580N1. Both configurations require a minimum of two vSAN disk groups per node and a minimum of three capacity disks per disk group. Depending on capacity requirements, you can configure up to four disk groups with a maximum of five capacity disks per disk group.
The VxRail D-Series nodes support up to eight disks per node. Only all-flash configurations are supported for SAP HANA for the dual-socket D560F with two vSAN disk groups per node and three capacity disks per disk group.
The VxRail E-Series nodes support up to 10 disks per node. Only all-flash and All-NVMe configurations are supported for SAP HANA for the dual-socket E560F1 and E560N1 with two vSAN disk groups per node and a minimum of three capacity disks per disk group.
Every VxRail node must have the same disk group configuration. For performance reasons, a best practice is to match the number of disk groups in the vSAN cluster to the number of production SAP HANA VMs running in the cluster.
The RAID 1 failure tolerance method is assigned in the storage policy to the SAP HANA persistence (the data and log of each SAP HANA VM), as explained in Assigning VM storage policies for SAP HANA data and log. RAID 1 failure tolerance creates a replica (mirroring) of the data and log devices and therefore consumes twice the usable SSD/NVMe capacity.
While RAID 1 is the default storage policy for all VMs on vSAN and is specifically required for production SAP HANA devices, you can create a new capacity-optimized storage policy with RAID 5/6 and assign it to other nonpersistent devices. RAID 5/6 consumes 1.33 times the usable device capacity on disk.