Infrastructure & Cloud
Understanding RAID for Small Businesses and Home Labs
Practical Examples Using an HP ProLiant DL380 Gen9
For many small businesses and home lab enthusiasts in the UK, a refurbished HP ProLiant DL380 Gen9 offers an excellent balance of performance, reliability, and affordability. These servers are widely available on the UK refurbished market and remain a popular platform for running virtual machines, file servers, backup systems, and learning enterprise infrastructure.
One of the most important storage decisions when deploying a server like this is choosing the correct RAID level. RAID determines how data is distributed across disks, how much usable storage you will have, and how resilient your system is if a drive fails.
In this article we will explore several common RAID configurations RAID 0, RAID 1, RAID 5, RAID 6, RAID 10, RAID 50, and RAID 60 using two realistic server configurations based on the HP ProLiant DL380 Gen9 platform.
The first configuration represents a typical entry-level or small business server with 8 × 900GB 2.5-inch SAS 10K drives, giving a total raw capacity of 7.2TB. The second configuration represents a larger, storage-focused deployment with 24 × 900GB 2.5-inch SAS 10K drives, providing 21.6TB of raw storage. These configurations reflect common setups used in virtualisation labs, business infrastructure, and storage servers.
RAID 0 – Performance Without Redundancy
RAID 0 is the simplest RAID configuration and focuses entirely on performance and storage capacity rather than data protection. In this setup, data is split into blocks and distributed evenly across all disks in the array, a process known as striping. Because multiple disks can read and write data simultaneously, RAID 0 offers excellent performance improvements compared with a single disk. On an HP ProLiant DL380 Gen9 equipped with 8 × 900GB SAS drives, RAID 0 would provide the full 7.2TB of usable storage, while a larger system with 24 drives would provide the entire 21.6TB. However, RAID 0 has a major drawback: it offers no redundancy whatsoever. If even a single disk in the array fails, the entire array becomes unusable and all data is lost because parts of every file are stored across multiple drives. For this reason, RAID 0 is rarely used for business-critical systems and is generally reserved for temporary workloads such as high-speed scratch storage, rendering workspaces, or non-critical lab environments where performance matters more than data safety.
RAID 1 – Mirroring for Simple Reliability
RAID 1 provides redundancy by mirroring data between two disks. Every piece of information written to one drive is duplicated on another drive, ensuring that if one disk fails, the system can continue operating using the remaining copy of the data. On a server such as the HP ProLiant DL380 Gen9, RAID 1 is most commonly used for operating system volumes rather than large storage pools. For example, two of the 900GB SAS drives might be configured as a RAID 1 mirror to host Windows Server, VMware ESXi, or a Linux installation. Because each disk contains identical data, only half of the total disk capacity is usable; two 900GB drives provide roughly 900GB of available storage. RAID 1 has the advantage of simplicity and reliability, and read performance can sometimes improve because the controller can retrieve data from either disk. However, its storage efficiency is relatively poor, making it impractical for large arrays of disks unless redundancy is the highest priority.
RAID 5 – Balanced Capacity and Fault Tolerance
RAID 5 is one of the most widely used RAID configurations in small business environments because it provides a good compromise between storage efficiency, redundancy, and performance. Instead of duplicating entire disks, RAID 5 distributes both data and parity information across all drives in the array. Parity data allows the system to reconstruct lost information if a single disk fails. With 8 × 900GB drives in a HP ProLiant DL380 Gen9, RAID 5 would use the equivalent capacity of one disk for parity, leaving approximately 6.3TB of usable storage. In a larger configuration with 24 drives, the array would provide roughly 20.7TB of usable capacity. RAID 5 offers good read performance and reasonable storage efficiency, but write performance is slightly reduced due to the need to calculate parity data. Additionally, rebuild times after a disk failure can be lengthy, and during this rebuild period the array is vulnerable to a second failure that could result in data loss. Despite these concerns, RAID 5 remains a practical option for moderately sized file servers or lab storage systems.
RAID 6 – Extra Protection for Larger Arrays
RAID 6 is similar to RAID 5 but improves reliability by using two independent parity calculations instead of one. This means the array can survive the failure of two disks simultaneously, which is particularly valuable when using large arrays with many drives. On an HP ProLiant DL380 Gen9 configured with 8 × 900GB SAS drives, RAID 6 would allocate the capacity of two disks to parity, resulting in approximately 5.4TB of usable storage. In the larger 24-drive configuration, RAID 6 would provide about 19.8TB of usable space. The additional parity improves fault tolerance but introduces a greater performance overhead for write operations because more parity calculations must be performed. RAID 6 is often recommended for storage arrays with large numbers of disks where the risk of multiple simultaneous failures increases, making it well suited to backup repositories, large file shares, and archival storage.
RAID 10 – High Performance with Strong Redundancy
RAID 10 (sometimes written as RAID 1+0) combines the advantages of RAID 1 mirroring with RAID 0 striping. In this configuration, disks are paired into mirrored sets, and data is then striped across those mirrored pairs. This design provides both redundancy and excellent performance because read and write operations can occur across multiple mirrored pairs simultaneously. With 8 × 900GB drives in a HP ProLiant DL380 Gen9, RAID 10 would create four mirrored pairs and stripe data across them, producing approximately 3.6TB of usable storage. In a larger 24-drive system, RAID 10 would provide roughly 10.8TB of usable capacity. Although the usable storage is only half of the total raw capacity, RAID 10 offers extremely fast performance and very quick rebuild times because only the failed disk’s mirror needs to be rebuilt. This makes RAID 10 an excellent choice for demanding workloads such as virtualisation platforms, databases, and high-IO applications commonly found in business servers or advanced home labs.
RAID 50 – Improved Performance for Large RAID 5 Arrays
RAID 50 combines the features of RAID 5 and RAID 0 by creating multiple RAID 5 groups and then striping data across them. This approach improves performance compared with a single large RAID 5 array and can also reduce the impact of a disk failure because only one subset of the array needs to rebuild. In a 24-disk configuration using the HP ProLiant DL380 Gen9, the drives might be divided into several RAID 5 groups which are then striped together. Depending on the exact layout, usable capacity would typically be around 20TB, while performance would be noticeably better than a standard RAID 5 configuration. RAID 50 is rarely used on very small arrays such as an 8-disk system because there are not enough drives to divide efficiently into multiple RAID 5 sets. However, in larger storage arrays it can provide a useful compromise between performance and capacity.
RAID 60 – High Redundancy for Enterprise Storage
RAID 60 works in a similar way to RAID 50 but uses RAID 6 groups instead of RAID 5 groups. Multiple RAID 6 arrays are created, each capable of surviving two disk failures, and then striped together using RAID 0. This design offers excellent fault tolerance because several disks can fail across different groups without data loss. In a 24-disk system based on the HP ProLiant DL380 Gen9, RAID 60 might be configured as two 12-disk RAID 6 arrays striped together, resulting in approximately 18TB of usable capacity. While RAID 60 provides exceptional reliability, it also requires a significant number of disks and introduces additional parity overhead, which reduces write performance compared with RAID 10. For most small business servers it is rarely necessary, but it can be valuable in large storage environments where maintaining uptime and protecting large datasets is the primary concern.
Choosing the Right RAID for a Small Business or Home Lab
For many small businesses or home lab environments running an HP ProLiant DL380 Gen9, the most practical RAID choices are RAID 10 for performance-sensitive workloads or RAID 6 for storage-focused systems where maximising capacity while maintaining redundancy is important. RAID 5 may still be used for smaller arrays, but modern storage practices often favour RAID 6 or RAID 10 due to the increased reliability they offer.
RAID 10 combines disk mirroring and striping, delivering excellent read and write performance while still providing redundancy. This makes it well suited for workloads such as virtual machines, databases, and transactional systems, where consistent I/O performance is critical. Because RAID 10 mirrors data across disk pairs, it can also provide faster rebuild times compared with parity-based RAID levels, reducing the amount of time the array remains in a degraded state.
RAID 6, on the other hand, is often chosen when usable storage capacity is a priority. By using dual-parity protection, RAID 6 can tolerate the failure of two disks simultaneously, which is particularly valuable when working with large arrays of enterprise drives such as 10K SAS disks. Although RAID 6 introduces a higher write penalty compared with RAID 10, it offers a strong balance between capacity efficiency and fault tolerance, making it suitable for file servers, backup repositories, and general business storage.
RAID 5 can still be a reasonable option in environments with smaller disk groups and moderate workloads, but its single-parity design means that arrays remain vulnerable during rebuild operations if another disk fails. As disk capacities increase and rebuild times become longer, many administrators now prefer RAID 6 or RAID 10 to reduce this risk.
Planning for RAID Rebuilds
Another important consideration when selecting a RAID level is how the array behaves during a rebuild. When a drive fails, the RAID controller must reconstruct the missing data onto a replacement disk using the remaining drives in the array. During this process the system operates in a degraded state, and performance can be significantly reduced depending on the RAID level and workload.
Parity-based arrays such as RAID 5 and RAID 6 typically experience longer rebuild times, especially with large disks or heavily utilised storage systems. RAID 10 arrays usually rebuild more quickly because only the mirrored pair needs to be reconstructed rather than recalculating parity across the entire array.
Understanding how rebuilds work and how to manage them safely is essential for anyone. For a detailed explanation of the rebuild process, monitoring rebuild progress, and best practices for replacing failed drives, you should read the RAID Rebuild Guide for the HP ProLiant DL380 Gen9, which walks through the full process and explains how to minimise risk while the array is recovering.
RAID Is Not a Backup
Regardless of the RAID configuration chosen, it is important to remember that RAID is not a backup solution. RAID protects against disk failures, but it does not protect against accidental deletion, ransomware, data corruption, or catastrophic hardware failure.
Small businesses and home lab operators should always maintain separate backups of critical data, ideally following the 3-2-1 backup strategy:
- 3 copies of your data
- 2 different storage media
- 1 off-site or cloud backup
By combining a well-planned RAID configuration with a reliable backup strategy, organisations can ensure both high availability and strong data protection for their server infrastructure.
