Data ONTAP takes extra steps to ensure that both disks in a carrier can be replaced without impacting any RAID group. Understanding this process helps you know what to expect when a disk from a multi-disk carrier storage shelf fails.
A multi-disk carrier storage shelf, such as the DS4486, has double the storage density of other SAS-connected storage shelves. It accomplishes this by housing two disks per disk carrier. When two disks share the same disk carrier, they must be removed and inserted together. This means that when one of the disks in a carrier needs to be replaced, the other disk in the carrier must also be replaced, even if it was not experiencing any issues.
Removing two data or parity disks from an aggregate at the same time is undesirable, because it could leave two RAID groups degraded, or one RAID group double-degraded. To avoid this situation, Data ONTAP initiates a storage evacuation operation for the carrier mate of the failed disk, as well as the usual reconstruction to replace the failed disk. The disk evacuation operation copies the contents of the carrier mate to a disk in a different carrier so that the data on that disk remains available when you remove the carrier. During the evacuation operation, the status for the disk being evacuated is shown as evacuating.
In addition, Data ONTAP tries to create an optimal layout that avoids having two carrier mates in the same RAID group. Depending on how the other disks are laid out, achieving the optimal layout can require as many as three consecutive disk evacuation operations. Depending on the size of the disks and the storage system load, each storage evacuation operation could take several hours, so the entire swapping process could take an entire day or more.
If insufficient spares are available to support the swapping operation, Data ONTAP issues a warning and waits to perform the swap until you provide enough spares.