Predictive disk failure and Rapid RAID Recovery

How Data ONTAP handles failing disks

Data ONTAP monitors disk performance so that when certain conditions occur, it can predict that a disk is likely to fail. For example, under some circumstances, if 100 or more media errors occur on a disk in a one-week period, Data ONTAP identifies this disk as a possible risk. When this occurs, Data ONTAP implements a process called Rapid RAID Recovery, and performs the following tasks:

Places the disk in question in pre-fail mode. This can occur at any time, regardless of what state the RAID group containing the disk is in.
Swaps in the spare replacement disk.
Copies the pre-failed disk's contents to a hot spare disk on the storage system before an actual failure occurs.
Once the copy is complete, fails the disk that is in pre-fail mode.

Steps 2 through 4 can only occur when the RAID group is in a normal state.

By executing a copy, fail, and disk swap operation on a disk that is predicted to fail, Data ONTAP avoids three problems that a sudden disk failure and subsequent RAID reconstruction process involves:

Rebuild time
Performance degradation
Potential data loss due to additional disk failure during reconstruction

If the disk that is in pre-fail mode fails on its own before copying to a hot spare disk is complete, Data ONTAP starts the normal RAID reconstruction process.