vSAN integrates storage parameters into vCenter Server by using vSphere APIs for storage awareness. SPBM further helps in defining VM-centric policies, which are basically constructs that store VM storage provisioning needs based on the available storage features with various policies (with configuration parameters).

These are given here:

• Number of disk stripes per object (performance): Default value = 01, maximum value = 12

Remarks: VM disk performance gets enhanced with RAID 0 stripe configuration by defining the HDD's number.

• Flash read cache reservation (%) (performance): Default % = 0, maximum % = 100

Remarks: We can leverage this configuration exclusively for VMs, which must have read IOPS issues. This needs to be sorted out, but reservations should be not recommended as per VMware best practices.

• Number of failures to tolerate, FTT (redundancy): Default value= 01, maximum value = 03

Remarks: The FTT number decides the number of host, disk, or network failures a storage object can tolerate. We can tolerate n (0, 1, 2, 3) failures when n+1 copies of the disk are created and 2n+1 hosts or fault domains need to contribute to the storage alongside mirroring. We can tolerate one failure, with minimum four nodes or fault domains and subsequent can tolerate two failures with minimum six hosts or fault domains along with erasure coding. The maximum value is 1 if the disk size is greater than 16 TB.

These parameters are configured in the Configure interface:

• Failure tolerance method (Performance/Capacity): Default = RAID 1 (Mirroring)-Performance

Remarks: RAID 1 can tackle failure tolerance using mirrors with good performance, while RAID 5/6 helps with failure tolerance by using parity blocks with great space efficiency. RAID 5/6 is only available on all-flash vSAN clusters, and when the Number of failures to tolerate (FTT) is set to 1 or 2. A value of 1 FTT implies a RAID 5 configuration, and a value of 2 FTT implies a RAID 6 configuration.

• IOPS limit for object (performance): Default = 0

Remarks: The IOPS limit for a disk IOPS is calculated as the number of I/Os using a defined size. It uses a base size of 32 KB by default, so a 64 KB I/O will represent 2 I/O. No limit policy is defined by setting the limit equal to 0.

• Disable object checksum (override policy): Default = No

Remarks: This setting determines whether checksums will be calculated for the data being written to the volume or not. Checksum calculation and error-correction are executed in the background. 

• Force provisioning (override policy): Default = No

Remarks: Force provisioning overrides the current policy if it won't comply with the available resources.

• Object space reservation (thick provisioning %): Default value = 0, maximum = 100

Remarks: It will help with a percentage of thick-provisioned storage objects upon VM creation, while the rest of the storage objects are thin-provisioned. When the expected amount of storage is already filled with objects, then it will help in reducing repeatable disk growth operation tasks.