Fractional reserve, also called LUN overwrite reserve, enables you to turn off overwrite reserve for space-reserved LUNs and files in a FlexVol volume. This can help you maximize your storage utilization, but if your environment is negatively affected by write operations failing due to lack of space, you must understand the requirements that this configuration imposes.

The fractional reserve setting is expressed as a percentage; the only valid values are 0 and 100 percent. The fractional reserve setting is an attribute of the volume.

Setting fractional reserve to 0 increases your storage utilization. However, an application accessing data residing in the volume could experience a data outage if the volume is out of free space, even with the volume guarantee set to volume. With proper volume configuration and use, however, you can minimize the chance of writes failing. ONTAP provides a “best effort” write guarantee for volumes with fractional reserve set to 0 when all of the following requirements are met:

Deduplication is not in use
Compression is not in use
FlexClone sub-files are not in use
All FlexClone files and FlexClone LUNs are enabled for automatic deletion

This is not the default setting. You must explicitly enable automatic deletion, either at creation time or by modifying the FlexClone file or FlexClone LUN after it is created.
ODX and FlexClone copy offload are not in use
Volume guarantee is set to volume
File or LUN space reservation is enabled
Volume Snapshot reserve is set to 0
Volume Snapshot copy automatic deletion is enabled with a commitment level of destroy, a destroy list of lun_clone,vol_clone,cifs_share,file_clone,sfsr, and a trigger of volume

This setting also ensures that FlexClone files and FlexClone LUNs are deleted when necessary.

If all the above requirements are met but your rate of change is high, in rare cases, the Snapshot copy automatic deletion could fall behind, resulting in the volume running out of space.
If all the above requirements are met and Snapshot copies are not in use, volume writes are guaranteed to not run out of space.

In addition, you can optionally use the volume autogrow capability to decrease the likelihood of volume Snapshot copies needing to be deleted automatically. If you enable the autogrow capability, you must monitor the free space in the associated aggregate. If the aggregate becomes full enough that the volume is prevented from growing, more Snapshot copies will probably be deleted as the free space in the volume is depleted.

If you cannot meet all of the above configuration requirements and you need to ensure that the volume does not run out of space, you must set the volume’s fractional reserve setting to 100. This requires more free space up front, but guarantees that data modification operations will succeed even when the technologies listed above are in use.

The default value and allowed values for the fractional reserve setting depend on the guarantee of the volume:

Volume guarantee	Default fractional reserve	Allowed values
Volume	100	0, 100
None	0	0, 100