The number of disks in a RAID (Redundant Array of Independent Disks) array can vary depending on the specific RAID level being used. RAID is a data storage technology that combines multiple disk drive components into a logical unit for the purposes of data redundancy, performance improvement, or both.
What is RAID?
RAID is an acronym that stands for Redundant Array of Independent Disks. It is a data storage technology that combines multiple physical disk drive components into one or more logical units for the purposes of data redundancy, performance improvement, or both.
Data is distributed across the drives in one of several ways, referred to as RAID levels, depending on the required level of redundancy and performance. The different schemes, or data distribution layouts, are named by the word RAID followed by a number, for example RAID 0 or RAID 1.
Each RAID level provides a different balance between these key goals: reliability and availability, performance, and capacity. Reliability and availability are achieved through data redundancy, where data is duplicated on more than one drive. This allows the array to remain functional even if one drive fails. Higher performance can come from distributing data across multiple drives, allowing them to operate in parallel. Capacity can be increased by combining multiple drives into a larger storage pool.
Why Use RAID?
There are several key benefits to using RAID technology:
- Increased data reliability and availability – By duplicating data across multiple drives, RAID can protect against data loss in the event of a drive failure.
- Improved performance – RAID allows drives to operate in parallel, increasing read and write speeds for higher overall throughput.
- Larger capacities – Multiple drives can be combined into a single logical volume, providing more storage capacity than a single drive.
- Flexibility – Different RAID levels allow the administrator to balance key factors like reliability, performance, and capacity.
For these reasons, RAID is commonly used in servers, networked storage devices, and other mission-critical computing systems where high availability, fast performance, and fault tolerance are important.
RAID Levels
There are several standard RAID levels, each with specific configurations in terms of drive utilization and data distribution:
RAID 0
- Data is striped across multiple drives with no parity or duplication.
- Best performance but no redundancy.
- At least 2 drives required.
RAID 1
- Drives are mirrored, with data duplicated on a second drive.
- Provides full redundancy but with limited performance gain.
- Minimum 2 drives required.
RAID 5
- Data is striped across drives with distributed parity allowing single drive failure tolerance.
- Good balance of performance and redundancy.
- Minimum 3 drives required.
RAID 6
- Similar to RAID 5 but with double distributed parity, allowing two drive failures.
- Less performance than RAID 5 but higher fault tolerance.
- Minimum 4 drives required.
RAID 10
- Combines mirroring (RAID 1) and striping (RAID 0) for both speed and redundancy.
- Fast performance and can survive multiple drive failures – as long as no mirror loses all drives.
- Minimum 4 drives required.
There are additional nested and non-standard RAID levels, but these encompass the most common configurations.
How Many Drives for Each RAID Level
The number of drives required for each standard RAID level is summarized below:
| RAID Level | Minimum Drives |
|---|---|
| RAID 0 | 2 |
| RAID 1 | 2 |
| RAID 5 | 3 |
| RAID 6 | 4 |
| RAID 10 | 4 |
Beyond the minimums, larger numbers of drives can be used in most RAID configurations for increased storage capacity and performance.
RAID 0
RAID 0 requires a minimum of 2 drives, but there is no upper limit. Two drives is the minimum needed to achieve data striping across multiple disks. Adding more disks increases overall storage capacity and can improve performance by allowing greater parallelism.
RAID 1
The minimum number of drives for RAID 1 is 2, providing a simple mirrored pair. For additional redundancy, RAID 1+0 (or RAID 10) combines mirroring with striping by creating multiple mirrored pairs. This can use 4, 6, 8, or more drives in increments of 2.
RAID 5
RAID 5 requires a minimum of 3 drives – this allows data striping across two drives with parity on the third drive. Additional drives can be added to grow overall capacity. Performance also increases up to a point, but becomes constrained by the parity drive overhead beyond 8-12 drives.
RAID 6
RAID 6 starts at 4 drives to allow striping with double distributed parity. Like RAID 5, additional drives can be added up to around 8-16 before performance becomes too limited by parity overhead.
RAID 10
RAID 10 is often configured using 4 drives, which allows two mirrored pairs using RAID 1+0. This can be expanded to use 6, 8, or more drives in increments of 2 to add more mirrored pairs.
Choosing the Right RAID Level
There are several factors to consider when choosing the appropriate RAID level:
- Performance needs – RAID 0 provides the best speed and throughput, while levels with parity introduce overhead.
- Redundancy requirements – Parity RAID levels like 5 and 6 provide single or double disk fault tolerance.
- Capacity requirements – RAID 0, 5, and 6 provide maximum overall storage capacity for the number of drives.
- Number of drives available – Each RAID level has minimum drive requirements.
- Availability and recovery time – Mirroring in RAID 1 and 10 allows for higher availability and faster rebuilding.
Here are some general guidelines and typical use cases for each RAID level:
- RAID 0 – Maximum performance with minimal disks. Used when speed is critical but redundancy is less important.
- RAID 1 – Mirroring for enhanced redundancy and availability. Used for small databases or other applications requiring fault tolerance.
- RAID 5 – Cost-effective storage redundancy. Widely used in network servers and storage arrays.
- RAID 6 – Double parity for highly critical data. Used in applications where RAID 5 may not provide enough fault tolerance.
- RAID 10 – Combines performance and redundancy for mission-critical systems. Popular for high demand databases and virtualization.
RAID Array Management
Managing and monitoring a RAID array is important to ensure protection and performance. Key aspects include:
- Checking array status regularly for any issues
- Reviewing drive health metrics for signs of potential failure
- Following best practices for drive firmware and RAID controller updates
- Monitoring drive utilization to watch for bottlenecks
- Planning for drive replacements before multiple failures occur
If a drive does fail, the impact and recovery process depends on the RAID level:
- RAID 0 – The entire array fails if any one drive is lost. All data is inaccessible until the failed drive is replaced and data is restored from backups.
- RAID 1 – If one drive fails, operations continue seamlessly on the surviving mirror. The failed drive can be replaced and rebuilt with data from the mirror.
- RAID 5 – The array remains operational if a single drive is lost. The missing data can be recreated from parity allowing the failed drive to be replaced without downtime.
- RAID 6 – Can withstand up to two drive failures. Failed drives are swapped out and rebuilt using parity data.
- RAID 10 – As long as one drive in each mirrored pair survives, the array stays active. Failed drives can be replaced and mirrored data recopied.
RAID Controller Considerations
A RAID controller is the hardware device that manages the array, including: striping and mirroring data across drives, calculating and writing parity, and handling recovery from drive failures.
Choosing an appropriate RAID controller is vital for optimal performance and reliability. Key factors include:
- RAID levels supported – The controller must support the required RAID level, such as RAID 5, 6, 10, etc.
- Drive interface and speed – Look for controllers with interfaces (SAS, SATA, etc.) that match the HDD or SSD drives being used in the array.
- Cache memory – More cache improves read/write speeds and allows the controller to manage a larger number of drives.
- Processing power – A faster CPU manages parity calculations more efficiently and improves overall throughput.
- Battery backup for cache – Protects against data loss in case of power failure.
- Redundant controllers – Providing failover in case the main controller goes down.
- Management software – Monitoring capabilities and ease of administration.
Conclusion
The number of disks in a RAID array depends primarily on the specific RAID level being implemented. Lower level RAIDs like 0, 1, and 5 have a minimum number of drives required to achieve their data layout, while also allowing drive counts to be scaled upwards for added capacity and performance.
Higher RAID levels like 6 and 10 provide specific combinations of redundancy and striping by using set quantities of mirrored and striped drives. For maximum performance with fault tolerance, RAID 10 is often implemented with larger quantities of drives, such as 8 or more.
When planning a RAID implementation, key considerations include the availability, capacity and performance needs of the application or system, in order to select both the appropriate RAID level and the optimal number of drives.