What is the RAID technology?
RAID (redundant array of independent disks) is a data storage virtualization technology that enables multiple drives of a storage to be seen as a unique logical device. The main purpose of this technology is to provide performance improvement and/or data redundancy. A RAID array of drives can be handled on the hardware as well as on the software level.
The DDP uses Areca RAID cards to manage its RAID arrays, which means it uses a hardware level RAID. By having a dedicated RAID management piece of hardware, the DDP software and CPU resources can focus on its main functions and are not affected by any RAID management load.
The way data is distributed across the drives is determined by the RAID level.
The RAID level that should be used depends on the level of redundancy and performance required.
The DDP typically uses RAID 5 and RAID 6 levels as it provides a protection against read errors, as well as whole disks failure. With RAID 5, one drive breakdown is tolerated. With RAID 6 up to two drives failures can be tolerated without any data loss. DDP can also be configured with hot spare drives.
When one drive fails, it doesn’t mean that it is necessarily faulty. As any computer, it can happen that a drive crashes for some reason. As you wouldn’t throw straight away your computer for a single crash, you could just insert the same disk to the DDP instead of replacing it.
When the drive is inserted back, the data rebuilding process automatically starts. If it can rebuild normally without any problem, it means that the drive just crashed for some reason.
The DDP mainly uses RAID 5 as it has been proven over the years to be reliable enough.
For an extra security, RAID 6 can be configured as well as hot spare drives so that the RAID arrays can be automatically rebuilt in case of drive failure.
The DDP is usually configured with 8 drives for each RAID 5 set. The capacity of 7 drives is available, 1 drive being dedicated to the parity.
For DDPs configured with RAID 6 sets, the capacity of 6 drives is available, 2 drives being dedicated to the parity.
Related to the RAID data protection mechanism, it is built around the parity concept.
What is the parity concept?
Widely used in the hardware industry, parity is used as an error detection technique and as a way to recover original data when a process fails to access it the usual way.
What is a read error?
On a physical disk, read errors can occur when the process that physically checks a disk block fails for some reason. In such a situation the parity block corresponding to the data that needs to be accessed can be read and the original data is recovered.
Hence when a read error is reported by the RAID card, it means that the parity data needed to be read in order to get the original data.