FAQ: Data, Parity, and Hot Spare Disks?
Data ONTAP allows disks to play different roles:
Every Redundant Array of Independent Disks (RAID) group requires one parity disk. The parity disk must be the same size or larger than the largest data disk (refer article 3011235: Does Data ONTAP allow mixing of drive sizes in a RAID group?). The parity disk is used in conjunction with the data disks to reconstruct data if a data disk fails. If the parity disk fails, the filer can reconstruct it from the data disks. With a single parity disk (RAID 4) setup, the system can tolerate a single disk failure per RAID group.
The double parity (dparity) disk will appear in RAID-DP groups. By default, aggregates created under Data ONTAP 7G are created RAID-DP (see the Storage Management Guide for more detail on the benefits of RAID-DP). The dparity disk allows the system to tolerate two disks failing in a single RAID group. If a dparity disk fails, the filer can reconstruct it from the existing data disks and/or parity.
Most disks in a typical filer are data disks. These store the filesystem and user data.
It is recommended to have at least one hot spare disk per filer. A hot spare is a disk that is not part of the RAID disk array and does not contain any file system data. After a disk failure, the filer will automatically begin to rebuild the file system onto the hot spare. Keep in mind that the hot spare must be equal or greater in size than the failed disk.
If there are no hot spares available, the filer will automatically shut down after 24 hours. This option can be changed to a higher value by changing the option raid.timeout, where the argument is the number of hours to wait before shutdown.
While Data ONTAP allows you to change the timeout option, running the filer without spares while one or more file system disks is failed (degraded mode) is not recommended.