When one of the hard drives in a RAID group fails and an idle or standby drive immediately replaces the failed hard drive, this drive is known as a “hot spare”.
The “hot spare” hard drive doesn’t store any data when it’s idle. It can be a global hot spare or a dedicated hot spare in a specific RAID group. When one of disks fails in a RAID group, the hot spare disk will replace the faulty one, and data will be rebuilt to the hot spare. This is only done via the data redundancy provided by other RAID disks, so RAID 0 cannot support hot spares. After finishing data rebuilding, a system administrator can take out the faulty disk, insert a new one, and assign it to be the new hot spare disk.
Suggestions for hot spares:
- You have to check that the drive supports hot sparing.
- The capacity must larger or equal to the faulty drive.
- Use the same drives (brand, specifications, speed, etc.) if possible. For example, when you use 3 SAS 15000 RPM hard drives in a RAID 5, it’s better that the hot spare is SAS 15000 RPM, too. If you use a SATA III, it will slow down the entire RAID’s performance.
RAID 5 + Hot Spare & RAID 6, which one is better？
RAID 5+Hot Spare: (N – 2) x (min. HDD capacity)
You need at least four drives. Three are for RAID 5, and one acts as a hot spare. (Note: a RAID group can have one hot spare only.)
RAID 6: (N – 2) x (min. HDD capacity)
You need at least 4 drives. It offers 2 hard drive redundancy, with data striped across multiple disks along with a parity check bits. The parity check bits ensure data integrity.
Both combinations share same number of drives and capacity, but RAID 6 offers a higher level data redundancy than RAID 5 and handles fault tolerance better, too. The overall data security for RAID 6 is better. The reason for this is that RAID 5 can only withstand one drive failure (with the hot spare drive ready to replace it). When there are two drive failures at the same time, the data cannot be saved. RAID 6, on the other hand, can withstand 2 drives failures at the same time. It does need time to rebuild data, however, on hot spare drive and this depends on the total capacity of the drives and the data’s size. There’s one con in regards to RAID 6: Because of the RAID 6’s additional fault tolerance mechanism, its performance is weaker than that of RAID 5.
We usually use the same batch of drives, so once one of the drives fail, it’s possible that other drives are going to fail, too. Taking this into account, there’s the risk that on RAID5 systems with hot spares that another drive will fail while in the process of rebuilding data.
Therefore, we should keep in mind that RAID data redundancy is not to the same as data backup. To ensure data safety, we need to back up our data daily to offline or off-site storage.
Please refer to “Remote Backup“.