RAID 5 is a commonly used RAID level that provides a good balance of storage capacity, performance, and redundancy for most applications. Determining the right number of drives for a RAID 5 setup depends on several factors.
What is RAID 5?
RAID 5 is a storage technology that combines multiple physical disk drives into one large logical drive. Data is striped across all the member drives, similar to RAID 0, but RAID 5 also utilizes distributed parity information to provide fault tolerance.
Specifically, RAID 5 writes parity information distributed across all member drives. If one physical drive fails, the parity information on the remaining drives can be used to reconstruct the data on the failed drive. This provides redundancy and protection against data loss.
A minimum of 3 drives are required for RAID 5. Common RAID 5 implementations include 3, 4, 5, or more drives. Larger RAID 5 arrays can sustain multiple drive failures, depending on the configuration.
Benefits of RAID 5
Compared to RAID 1 mirroring, RAID 5 provides better storage efficiency as the available capacity is n-1 drives instead of half the number of drives. For example, four 2TB drives in RAID 5 would yield 6TB of usable capacity, while RAID 1 would provide only 4TB from the same number of disks.
RAID 5 performance is better for read operations than RAID 1 or RAID 10 since the load can be distributed across all member disks. Writes are slower due to the parity calculation overhead, making RAID 5 more suitable for read-intensive workloads.
The distributed parity implementation in RAID 5 eliminates the write bottleneck that exists with RAID 4 parity. RAID 5 is generally more preferred over RAID 4.
By providing redundancy, RAID 5 protects against data loss from a single drive failure. Additional spare drives can allow recovery from multiple drive failures.
Disadvantages of RAID 5
While RAID 5 provides good redundancy for most use cases, it lacks the fault tolerance of mirroring configurations like RAID 1 or RAID 10. With RAID 5, a single drive failure results in reduced redundancy until the failed drive is replaced.
RAID 5 write performance and fault tolerance are reduced during drive rebuilds. The controller must recalculate parity across all drives when rebuilding, increasing load during the rebuild operation.
Larger RAID 5 implementations also have greater risk of unrecoverable read errors during rebuilds. With more drives, there is a higher likelihood of data loss in case of multiple drive failures exceeding parity protections.
For these reasons, RAID 6 is sometimes used instead of RAID 5 for the added redundancy of dual distributed parity on very large arrays.
Minimum Number of Drives
The minimum number of drives required for RAID 5 is 3. This provides 1 drive worth of redundancy in the array.
With a 3-drive RAID 5 array:
- Total capacity is size of 2 drives
- Can sustain 1 drive failure without data loss
While functional, the 3-drive minimum RAID 5 configuration lacks expandability. It also provides the lowest storage efficiency, with only 66% of total capacity as usable space.
4 Drives
4-drive RAID 5 arrays are a popular choice for many use cases. Benefits include:
- Total capacity is size of 3 drives
- Can sustain 1 drive failure without data loss
- Efficiency of 75% usable capacity
- Good performance with load balancing across 4 drives
- Low cost expandability by adding drives in pairs
The 4-drive RAID 5 configuration provides a good balance of capacity, redundancy, and cost for general purpose storage needs. It’s a flexible option allowing for future expansion.
6 Drives
A 6-drive RAID 5 array provides:
- Total capacity of 5 drives
- Can sustain up to 2 drive failures
- 83% efficiency
The increased redundancy makes 6-drive RAID 5 a good choice for more demanding applications where maximum uptime is critical. The two drive fault tolerance provides safety from double drive failures that may occur during rebuilds.
8 Drives
The benefits of an 8-drive RAID 5 implementation are:
- Capacity of 7 drives
- Up to 3 drive fault tolerance
- 87% efficiency
As drive counts go higher, the likelihood of multiple failures rises. 8-drive configurations provide strong protection by allowing continuous operations even after 2 or 3 spindles fail.
Larger Implementations
Larger RAID 5 arrays with 10+ drives are less common outside of enterprise environments. Benefits over smaller arrays include:
- Higher capacities and disk count improve performance
- Increased redundancy against drive failures
- Up to 3 spare drives can be provisioned for auto-rebuilding failed disks
However, very large RAID 5 arrays also have drawbacks:
- High risk of unrecoverable reads during rebuilds
- Degraded performance during rebuilds due to increased load
- Higher likelihood of multiple drive failures exceeding parity protections
For these reasons, RAID 6 dual parity or RAID 10 mirroring may be better options for large storage pools where maximum performance and fault tolerance are critical.
RAID 5 Storage Efficiency
The table below shows the total capacity, usable capacity, and storage efficiency of sample RAID 5 configurations:
| Drives | Total Capacity | Usable Capacity | Efficiency |
|---|---|---|---|
| 3 x 2TB | 6TB | 4TB | 66% |
| 4 x 2TB | 8TB | 6TB | 75% |
| 6 x 2TB | 12TB | 10TB | 83% |
| 8 x 2TB | 16TB | 14TB | 87% |
As drive counts scale up, RAID 5 usable capacity gets closer to total raw capacity. But storage efficiency declines again with very large arrays as more spare drives are required.
Performance Considerations
RAID 5 performance is influenced by the number of drives in the array. In general:
- More drives provide faster read speeds by distributing load
- However, write speeds will suffer due to parity calculation overhead
- Rebuild times are slower with more drives to reconstruct
Workloads with heavy write activity may benefit from a RAID 10 mirroring configuration instead of RAID 5 for better performance.
Enterprise RAID controllers and SSDs can mitigate some of the RAID 5 write penalty issues for applications requiring faster write speeds.
Expandability
Larger RAID 5 implementations allow for more incremental expandability. For example:
- 3 drive RAID 5 cannot be expanded without full rebuild
- 4 drive RAID 5 can be expanded by adding pairs of disks
- Higher drive counts can add multiple disks at a time
The ability to expand storage by adding disks is useful for cost-effective scaling as data storage needs grow.
Conclusion
RAID 5 drive requirements depend on the specific needs for capacity, redundancy, performance, and future expansion.
For home or small office use, 4 to 6 drive RAID 5 arrays provide a good balance of storage efficiency, fault tolerance, and low cost. This makes RAID 5 ideal for general purpose file and application servers.
Larger RAID 5 implementations with 8 to 12 drives work well for mid-range applications that demand higher capacity, I/O performance, and availability. Media streaming, databases, and other read-intensive workloads are good fits for larger RAID 5 storage pools.
Very large RAID 5 installations are only recommended when using high end RAID controllers and SSDs to improve performance and rebuild times. For most high capacity enterprise storage needs, RAID 6 or RAID 10 scale better while providing increased redundancy and fault tolerance.