This chapter covers RAID levels 0, 1, 5, 6, and 10 — the most common configurations tested on the CompTIA A+ 220-1101 exam. Understanding RAID is essential for storage troubleshooting, system configuration, and data redundancy decisions. RAID questions appear on roughly 5-8% of the exam, often in scenario-based questions where you must select the appropriate RAID level for performance, redundancy, or capacity requirements.
Jump to a section
Imagine a busy office where employees need to store and retrieve important documents. RAID 0 is like having two filing cabinets but storing half of each document in each cabinet with no duplication. If one cabinet is destroyed, you lose half of every document — no redundancy. RAID 1 is like having a photocopier that makes an exact copy of every document and stores it in a second cabinet. If one cabinet is lost, you have a perfect copy. RAID 5 is like having three cabinets where each document is split into two parts, and a 'parity' page is created that mathematically represents the missing part. If one cabinet is lost, you can use the remaining parts and the parity page to reconstruct the missing piece. It's efficient but requires calculation. RAID 6 is like RAID 5 but with two different parity pages stored in separate cabinets, so you can lose any two cabinets and still reconstruct all data. RAID 10 is like having two pairs of mirrored cabinets, and then spreading documents across both pairs for speed. You can lose one cabinet from each pair without data loss, but losing both cabinets in a pair is catastrophic. Each level trades off between speed, capacity, and fault tolerance.
What is RAID and Why Does It Exist?
RAID (Redundant Array of Independent Disks) combines multiple physical disk drives into a single logical unit to improve performance, reliability, or both. The key concepts are:
Striping: Splitting data across multiple disks to increase throughput.
Mirroring: Duplicating data across multiple disks for redundancy.
Parity: Using mathematical calculations (XOR) to reconstruct data if a disk fails.
RAID can be implemented via hardware (a dedicated RAID controller) or software (OS-level, e.g., Windows Storage Spaces, Linux mdadm). For the 220-1101 exam, focus on the characteristics of each level — not the implementation details.
RAID 0 – Striping
RAID 0 splits data into blocks and writes them across two or more disks in a stripe. No redundancy exists; if one disk fails, all data is lost.
Minimum disks: 2
Capacity: 100% of total raw capacity (e.g., 2 × 1 TB = 2 TB usable)
Performance: Excellent read/write speeds because multiple disks work in parallel.
Fault tolerance: None. Failure of any disk destroys the array.
Use case: Temporary data, scratch disks, or applications where speed is critical and data loss is acceptable (e.g., video editing cache).
Exam tip: RAID 0 is the only level that increases usable capacity without any overhead, but it offers zero redundancy.
RAID 1 – Mirroring
RAID 1 writes identical data to two or more disks simultaneously. If one disk fails, the system continues using the other.
Minimum disks: 2
Capacity: 50% of total raw capacity (e.g., 2 × 1 TB = 1 TB usable)
Performance: Read speeds can improve (since both disks can serve reads), but write speed is limited by the slowest disk.
Fault tolerance: Can survive failure of one disk (in a 2-disk array) or up to n-1 disks if more than two are mirrored.
Use case: Operating system drives, critical databases, or any scenario where data availability is paramount.
Common wrong answer: Candidates think RAID 1 doubles read speed. While reads can be faster if the controller supports simultaneous reads from both disks, write speed is not improved.
RAID 5 – Striping with Distributed Parity
RAID 5 stripes data across all disks and stores parity information (an XOR calculation of the data blocks) distributed across the drives. If one disk fails, the parity from the remaining disks can reconstruct the missing data.
Minimum disks: 3
Capacity: (n-1) / n of total raw capacity (e.g., 3 × 1 TB = 2 TB usable; 4 × 1 TB = 3 TB usable)
Performance: Good read speeds; write speeds are slower due to parity calculation overhead.
Fault tolerance: Can survive one disk failure. During rebuild, performance degrades significantly, and a second failure during rebuild causes data loss.
Use case: File servers, general-purpose storage where a balance of capacity, performance, and redundancy is needed.
Exam trap: The parity is *distributed* across all disks — not stored on a dedicated parity disk. This avoids a write bottleneck on a single disk.
RAID 6 – Striping with Double Parity
RAID 6 is similar to RAID 5 but uses two independent parity blocks (using different algorithms) distributed across the array. It can survive two simultaneous disk failures.
Minimum disks: 4
Capacity: (n-2) / n of total raw capacity (e.g., 4 × 1 TB = 2 TB usable; 5 × 1 TB = 3 TB usable)
Performance: Read speeds similar to RAID 5; write speeds are slower due to double parity calculation.
Fault tolerance: Can survive two disk failures. Rebuild times are longer than RAID 5.
Use case: Large-capacity storage arrays where rebuild times are long and the risk of a second failure during rebuild is unacceptable (e.g., archival storage, large databases).
Exam tip: The exam may ask which RAID level can survive two disk failures — the answer is RAID 6. Also note RAID 6 requires more disks for the same usable capacity compared to RAID 5.
RAID 10 (1+0) – Striping of Mirrors
RAID 10 combines mirroring and striping. First, disks are paired into mirrored sets (RAID 1), then those sets are striped together (RAID 0). This provides both redundancy and performance.
Minimum disks: 4 (must be even number)
Capacity: 50% of total raw capacity (e.g., 4 × 1 TB = 2 TB usable)
Performance: Excellent read and write speeds because striping increases throughput and mirroring allows read balancing.
Fault tolerance: Can survive multiple disk failures as long as no mirrored pair loses both disks. For example, in a 4-disk array (two mirrored pairs), you can lose one disk from each pair and still have all data.
Use case: High-performance databases, virtualization hosts, mission-critical applications requiring both speed and redundancy.
Common wrong answer: Candidates confuse RAID 10 with RAID 01 (mirroring of stripes). RAID 10 is far more common and is the one tested. RAID 01 is less fault-tolerant because a single disk failure in one stripe can cause the entire stripe set to be unavailable.
RAID Configuration and Verification Commands
Software RAID (Linux mdadm):
# Create a RAID 0 array from /dev/sdb and /dev/sdc
mdadm --create /dev/md0 --level=0 --raid-devices=2 /dev/sdb /dev/sdc
# Create a RAID 1 array
mdadm --create /dev/md1 --level=1 --raid-devices=2 /dev/sdd /dev/sde
# Create a RAID 5 array
mdadm --create /dev/md2 --level=5 --raid-devices=3 /dev/sdf /dev/sdg /dev/sdh
# Create a RAID 10 array
mdadm --create /dev/md3 --level=10 --raid-devices=4 /dev/sdi /dev/sdj /dev/sdk /dev/sdl
# Check RAID status
cat /proc/mdstat
mdadm --detail /dev/md0Windows Storage Spaces:
Use Server Manager or PowerShell to create storage pools and virtual disks with resiliency types: Simple (RAID 0), Mirror (RAID 1), Parity (RAID 5/6-like).
Hardware RAID controllers:
Boot into the RAID BIOS (e.g., Ctrl+R for Dell PERC, Ctrl+H for HP Smart Array) to create arrays.
Use vendor-specific tools like hpacucli or storcli for management.
How RAID Interacts with Other Technologies
SSD vs HDD: RAID with SSDs reduces the impact of parity calculation latency because SSDs have much lower access times. However, write endurance may be a concern for parity RAID.
Hot spares: A standby disk that automatically replaces a failed disk in the array. Supported by RAID 1, 5, 6, 10.
JBOD: Just a Bunch Of Disks — not RAID. Disks are presented individually. The exam may ask which is not a RAID level; JBOD is not.
RAID and backups: RAID is not a backup. It provides availability, not protection against accidental deletion, corruption, or disasters.
Trap Patterns on the Exam
RAID 0 vs RAID 1: Candidates often choose RAID 0 for redundancy because they associate "0" with "no loss" — actually the opposite.
RAID 5 minimum disks: Many think 2 disks for RAID 5 — it's 3.
RAID 10 vs RAID 01: The exam may describe a scenario where you need both speed and redundancy; RAID 10 is the correct answer.
Fault tolerance numbers: RAID 5 tolerates 1 failure; RAID 6 tolerates 2; RAID 1 (2-disk) tolerates 1; RAID 10 tolerates up to half the disks failing as long as they are in different mirrors.
Determine Requirements
Identify the primary goal: performance (RAID 0), redundancy (RAID 1), capacity with redundancy (RAID 5/6), or both performance and redundancy (RAID 10). Count the number of available disks and consider the acceptable capacity loss. For example, if you have 4 × 1 TB drives and need maximum performance with redundancy, RAID 10 gives 2 TB usable. If capacity is more important and you can tolerate one disk failure, RAID 5 would give 3 TB usable.
Choose RAID Level
Based on requirements, select the appropriate RAID level. For the exam, know the minimum disk count, capacity efficiency, fault tolerance, and typical use cases. For instance, a web server needing fast reads and writes with redundancy would likely use RAID 10. A large media archive with limited budget might use RAID 6 to protect against two drive failures during long rebuilds.
Prepare Disks
Ensure all disks are of the same size and speed for optimal performance. In RAID 5/6, the array capacity is limited by the smallest disk. Disks should be connected to the same controller (hardware) or be available as block devices (software). Initialize disks (e.g., using `fdisk` or Disk Management) to clear any existing partitions.
Create the Array
Use the appropriate tool: RAID BIOS for hardware controllers, or software tools like `mdadm` (Linux) or Storage Spaces (Windows). Specify the RAID level, number of devices, and the actual disk devices. For example, to create RAID 5 with three disks: `mdadm --create /dev/md0 --level=5 --raid-devices=3 /dev/sdb /dev/sdc /dev/sdd`. The array will synchronize (build parity) — this can take hours for large drives.
Format and Mount
After the array is created, create a filesystem (e.g., ext4, NTFS) on the logical device. Then mount it to a directory or assign a drive letter. Verify the mount and test read/write operations. For example: `mkfs.ext4 /dev/md0; mount /dev/md0 /mnt/raid`. Ensure the mount is persistent by adding an entry to /etc/fstab.
Enterprise Scenario 1: Virtualization Host Storage
A data center running 50 virtual machines on a single hypervisor host needs fast, redundant storage. The host has 8 × 1.2 TB SAS drives. The engineer configures RAID 10: four mirrored pairs striped together. This yields 4.8 TB usable, excellent IOPS for database VMs, and the ability to survive up to four disk failures (one per mirror). During a drive failure, the hot spare automatically rebuilds the mirror pair without downtime. The key performance consideration is that RAID 10 write performance scales with the number of spindles, so 8 drives give near-linear write throughput. Misconfiguration would be using RAID 5 with 8 drives — the parity calculation overhead would degrade write performance by 20-30%, and the rebuild time after a failure could exceed 24 hours, risking a second failure.
Enterprise Scenario 2: Large Archival Storage
A media company stores 500 TB of video files. They use a storage array with 60 × 10 TB HDDs. RAID 6 is chosen because the rebuild time for a 10 TB drive is over 12 hours, and the probability of a second failure during that window is non-trivial. RAID 6 allows two simultaneous disk failures. The usable capacity is 58 × 10 TB = 580 TB. Performance is not critical for archival access, so the write penalty of double parity is acceptable. The engineer configures hot spares (2 drives) and monitors the array for degraded states. A common mistake is using RAID 5 — a second failure during rebuild would cause total data loss. RAID 6 also helps when a URE (Unrecoverable Read Error) occurs during rebuild, as the second parity block can still reconstruct data.
Enterprise Scenario 3: High-Availability Database Server
A financial application requires a database with very high write throughput and no downtime. The server has 12 × 600 GB 15K RPM SAS drives. The engineer deploys RAID 10 with 6 mirrored pairs. This provides 3.6 TB usable, excellent random write performance (because writes go to all mirrors simultaneously), and the ability to lose up to 6 drives as long as no pair loses both. The database transaction logs are stored on a separate RAID 1 array (2 drives) for extra safety. A common pitfall is using RAID 5 for database logs — the write penalty causes transaction log bottlenecks. RAID 10 avoids this by not requiring parity calculations. The array is monitored using SMART data and the RAID controller's management software to predict failures.
Exactly What the 220-1101 Tests
Objective 3.4 (Storage) includes RAID levels. The exam expects you to:
Identify RAID levels 0, 1, 5, 6, and 10 by their characteristics.
Select the appropriate RAID level for a given scenario (performance, redundancy, capacity).
Know the minimum number of disks required for each level.
Understand fault tolerance: how many disk failures each level can survive.
Recognize that RAID is not a backup — it provides availability, not data protection against deletion or corruption.
Most Common Wrong Answers and Why Candidates Choose Them
Choosing RAID 0 for redundancy: Candidates see "0" and think it means no data loss. Actually RAID 0 has zero redundancy.
Choosing RAID 1 for capacity: RAID 1 uses half the capacity for mirroring. Candidates forget the 50% overhead.
Thinking RAID 5 needs 2 disks: Minimum is 3. The parity calculation requires at least three disks.
Confusing RAID 10 with RAID 01: RAID 10 (striped mirrors) is more fault-tolerant. RAID 01 (mirrored stripes) can lose data if a single disk in a stripe fails.
Specific Numbers and Terms That Appear on the Exam
RAID 0: minimum 2 disks, no fault tolerance, 100% capacity utilization.
RAID 1: minimum 2 disks, survives 1 failure, 50% capacity.
RAID 5: minimum 3 disks, survives 1 failure, (n-1)/n capacity.
RAID 6: minimum 4 disks, survives 2 failures, (n-2)/n capacity.
RAID 10: minimum 4 disks (even number), survives up to half the disks (if in different mirrors), 50% capacity.
Edge Cases and Exceptions
RAID 1 with more than 2 disks: Can survive n-1 failures. For example, a 3-disk RAID 1 can survive 2 failures.
RAID 5 with a hot spare: The hot spare is not part of the array until activated. The array still tolerates only one failure at a time.
RAID 6 write penalty: Each write requires two parity calculations, making it slower than RAID 5.
RAID 10 rebuild: If a disk fails, only its mirror pair is affected. Rebuild is faster than RAID 5/6 because only the mirror needs to be copied.
How to Eliminate Wrong Answers
If the scenario mentions "no redundancy" or "maximum performance," eliminate RAID 1, 5, 6, 10 (all have some redundancy).
If the scenario mentions "survive two disk failures," the only option among the five is RAID 6.
If the scenario requires both speed and redundancy with at least 4 disks, RAID 10 is likely correct.
If the scenario mentions "minimum cost" and "fault tolerance for one disk," RAID 5 is a good balance.
Remember that JBOD is not RAID — it's often a distractor.
RAID 0: striping, no redundancy, minimum 2 disks, 100% capacity, best performance.
RAID 1: mirroring, minimum 2 disks, 50% capacity, survives 1 failure.
RAID 5: striping with distributed parity, minimum 3 disks, (n-1)/n capacity, survives 1 failure.
RAID 6: striping with double parity, minimum 4 disks, (n-2)/n capacity, survives 2 failures.
RAID 10: striping of mirrors, minimum 4 disks (even), 50% capacity, survives multiple failures as long as no mirror pair loses both disks.
RAID is not a backup — it provides availability, not data protection against logical errors.
Hardware RAID uses a dedicated controller; software RAID uses OS resources.
Hot spares can automatically replace failed disks in hardware RAID.
These come up on the exam all the time. Here's how to tell them apart.
RAID 5
Minimum 3 disks
Capacity efficiency: (n-1)/n (e.g., 3 of 4 TB usable)
Write performance degraded due to parity calculation
Fault tolerance: 1 disk failure
Rebuild time longer due to parity reconstruction
RAID 10
Minimum 4 disks (even number)
Capacity efficiency: 50% (e.g., 2 of 4 TB usable)
Excellent write performance (no parity)
Fault tolerance: up to half the disks (if in different mirrors)
Faster rebuild (only mirror copy needed)
Mistake
RAID 0 provides data redundancy because the data is spread across multiple drives.
Correct
RAID 0 provides no redundancy. If any drive fails, all data is lost. The striping improves performance but not fault tolerance.
Mistake
RAID 5 can survive two disk failures.
Correct
RAID 5 can survive only one disk failure. RAID 6 is required for two simultaneous failures.
Mistake
RAID 10 requires exactly 4 disks.
Correct
RAID 10 requires a minimum of 4 disks, but can use any even number greater than 2 (e.g., 6, 8, 10).
Mistake
RAID 1 doubles write speed because data is written to two disks simultaneously.
Correct
RAID 1 writes are not faster — they are limited by the slowest disk. Read speed can improve if the controller reads from both disks concurrently.
Mistake
RAID is a substitute for backups.
Correct
RAID provides high availability against disk failure, but does not protect against data corruption, accidental deletion, or disasters. Backups are still required.
Reveal each answer, then mark whether you got it right. Score 60%+ to unlock the next chapter.
RAID 5 uses one parity block per stripe and can survive one disk failure. RAID 6 uses two parity blocks per stripe and can survive two simultaneous disk failures. RAID 6 requires a minimum of 4 disks (vs. 3 for RAID 5) and has lower write performance due to double parity calculations. For the exam, remember that RAID 6 is for environments where rebuild times are long and the risk of a second failure is high.
Technically yes, but it is not recommended because if any disk fails, the OS and all data are lost. RAID 1 is a better choice for OS drives to ensure availability. The exam may present a scenario where a user wants speed for a gaming system — RAID 0 could be acceptable for non-critical data, but not for an OS.
The minimum is 4 disks. RAID 10 requires at least two mirrored pairs (each pair is RAID 1) that are then striped (RAID 0). So you need 2 pairs × 2 disks each = 4 disks. The number must be even.
RAID 5 distributes parity across all disks. For each stripe, one block contains parity information (XOR of the data blocks in that stripe). The parity block rotates among disks to avoid a single disk becoming a bottleneck. If one disk fails, the controller uses the parity from the remaining disks to reconstruct the missing data.
In theory, RAID 0 can be slightly faster because it has no redundancy overhead, but RAID 10 also offers excellent performance because it stripes data across mirrors. In practice, RAID 10 often has comparable read speeds and may have better write performance due to no parity calculation. The exam expects you to know that RAID 0 has the best performance but no redundancy, while RAID 10 balances performance and redundancy.
The array enters a degraded state. Read performance may be slower because missing data must be reconstructed from parity. Write performance also degrades because parity must be recalculated. The system continues to operate. The failed disk must be replaced and the array rebuilt. During rebuild, the array is vulnerable — if another disk fails, all data is lost.
Yes, but the array capacity is limited by the smallest disk. For example, in a RAID 5 with 1 TB, 2 TB, and 3 TB disks, each disk is treated as 1 TB, so usable capacity is 2 TB (3 × 1 TB - 1 parity). The extra space on larger disks is wasted. For the exam, know that disk sizes should ideally match to avoid waste.
You've just covered RAID Levels: 0, 1, 5, 6, 10 — now see how well it sticks with free 220-1101 practice questions. Full explanations included, no account needed.
Done with this chapter?