Skip to main content
StashRaid

Storage Parameters

Usable Capacity

12 TB

Fault Tolerance

4

Array Resilience

Architecture

RAID 5 • Parity

Fault Tolerance

1 Concurrent Failures Allowed

Current Configuration

RAID 5 (Parity)

12 TB Usable

Total RAW: 16 TB

Disk 01DATA
Disk 02DATA
Disk 03DATA
Disk 04PARITY

Total Drives

4 x 4 TB

Efficiency

75% Utilized

Usable: 12 TBWasted to redundancy: 4 TB

Pros & Cons

  • PRO: Good balance of capacity/speed/safety

  • PRO: Cost-effective redundancy

  • CON: Slower writes due to parity calculation

    This setup prioritizes redundancy but has associated performance or capacity costs.

RAID Calculator

Estimate your usable storage, parity overhead, and drive loss tolerance instantly. Our tool helps you plan your storage array before buying hardware or initializing disks.

Compare RAID 0, 1, 5, 6, 10, SHR & ZFS Levels

Understanding the tradeoffs between capacity and fault tolerance is critical. For instance, RAID 0 offers maximum space but no protection, while RAID 1 mirrors everything. RAID 5 and RAID 6 use parity math for a middle ground. SHR (Synology Hybrid RAID) maximizes capacity with mixed drive sizes — only the largest drive is lost to redundancy. For data integrity, ZFS RAID-Z1/Z2/Z3 add checksumming and eliminate the write hole.

How Usable Capacity Is Calculated

Raw capacity represents the absolute sum of your drives, but usable capacity subtracts formatting overhead (base-2 math vs base-10 marketing) and the structural parity required for the chosen raid level. Learn more in our capacity guide.

RAID Calculator FAQ

What is a RAID calculator?

It's a tool that takes your number of hard drives, their sizes, and a requested RAID configuration to instantly output exactly how much space you will be able to use for real data.

Can I mix different drive sizes in a RAID array?

Traditional RAID levels force all drives to match the smallest disk, wasting space on larger drives. SHR (Synology Hybrid RAID) solves this by partitioning drives into slices and spanning multiple internal RAID arrays — only the largest drive is lost to redundancy. Unraid and ZFS also handle mixed drives differently: Unraid uses a dedicated parity drive (must be ≥ largest data drive), while ZFS vdevs still default to smallest-drive behavior.

Which RAID level is best for a NAS?

For modern arrays using huge drives (10TB+), RAID 6 is strongly recommended to protect against sector errors during a rebuild. For maximum performance over network, RAID 10 is preferred.

The Storage Architect's RAID Guide

Rule #1: RAID is NOT a Backup

This is the most critical concept in storage. RAID protects your server against hardware drive failure so that your services can keep running while a drive is replaced. It does absolutely nothing to protect against:

  • Accidental file deletion
  • Ransomware or malware encryption
  • Catastrophic hardware failure (e.g., PSU fries everything)
  • Fire, flood, or theft

Always implement the 3-2-1 backup strategy: 3 copies of your data, on 2 different media, with 1 copy off-site.

SHR & SHR-2: Synology Hybrid RAID

The Mixed-Drive Solution

SHR is Synology's automated RAID system that solves traditional RAID's biggest weakness: it allows mixing drives of different sizes without wasting capacity. Built on Linux mdadm and LVM, SHR partitions each drive into slices matching the smallest common denominator, creates multiple RAID arrays from those slices, and spans them into a single volume.

SHR (1-Drive Redundancy)

Equivalent to RAID 5 in fault tolerance. With identical drives, capacity = (N-1) × drive. With mixed drives, only the single largest drive is "lost" — remaining space is fully utilized.

  • • Start with just 2 drives, add more at any time without rebuilding.
  • • Mix any capacities — the system auto-optimizes parity distribution.
  • • Drives can be moved to another Synology NAS and recognized immediately.

SHR-2 (2-Drive Redundancy)

Equivalent to RAID 6 in fault tolerance. Two largest drives are "lost" to double parity. Recommended for arrays of 6+ drives or drives larger than 8TB.

  • • Survives 2 simultaneous drive failures while maintaining mixed-drive flexibility.
  • • Safe against UREs during rebuild — second parity layer provides fallback.
  • • Use the SHR calculator to estimate capacity.

Important SHR Limitation

SHR arrays can only be read by Synology NAS devices. If your Synology enclosure dies and you want to recover data by connecting drives to a PC, you'll need specialized recovery software (like UFS Explorer) that understands SHR's layered RAID structure. Standard RAID 5/6 or ZFS arrays can be imported by any Linux system with mdadm or ZFS support.

RAID Is NOT a Backup

RAID protects against drive failure — not accidental deletion, ransomware, fire, or theft. Even SHR-2 with dual-drive redundancy is not a substitute for proper backups. Always follow the 3-2-1 rule: 3 copies of your data, on 2 different media types, with 1 copy off-site.

How SHR Parity Protects All Your Data

New SHR users are often confused by Synology showing "21.8TB available, 7.3TB protection" for a 4-drive array. All 21.8TB is fully protected by that 7.3TB of parity space. Parity works like simple arithmetic: store the sum of your data drives on the parity drive, and any single missing value can be reconstructed. With 2 drives, SHR mirrors everything (50/50 split). With 3+ drives, it switches to parity — one drive's capacity protects all data drives, so efficiency rises from 50% to 75% to 87.5% as you add drives. Learn more on the SHR calculator page →

Software vs. Hardware RAID

Modern Software RAID (ZFS, Unraid, mdadm)

In the modern era, software RAID is strongly preferred. Systems like TrueNAS (using ZFS) manage the disks directly through the OS.

  • • Protects against silent data corruption (bit-rot) via checksumming.
  • • Hardware independent—move your drives to a new motherboard and it just works.
  • • Can utilize system RAM for massive caching (ARC).

Legacy Hardware RAID (RAID Cards)

Historically, RAID was calculated by a dedicated processor on a PCIe card to offload work from a weak CPU.

  • • Hardware lock-in: if the card dies, you need an identical card to read the array.
  • • Hides SMART data from the OS (usually).
  • • Vulnerable to the "Write Hole" parity bug during power failures unless equipped with a dead-battery-backed cache.

The URE Threat & Why RAID 5 is risky for large drives

When a drive fails in RAID 5, the array must read every single sector of the remaining drives to calculate the missing data and reconstruct it onto a new replacement drive.

Consumer large-capacity drives (e.g., 10TB+) have an Unrecoverable Read Error (URE) rate—typically 1 in 10^14 bits read. When rebuilding a massive array, the mathematical probability of encountering a URE before the rebuild completes becomes terrifyingly high. If a URE occurs during a RAID 5 rebuild, the rebuild fails, and the array collapses.

Architecture Recommendation

For arrays using drives larger than 8TB, always use RAID 6 (Dual Parity/RAIDZ2) or RAID 10 (Striped Mirrors). RAID 6 can survive a URE during a rebuild because it has a second layer of parity to fall back on.

Homelab Hardware Guide

Drive Technology: CMR vs SMR

Never put a Shingled Magnetic Recording (SMR) drive in a RAID array. They overlap data tracks to increase density, which causes catastrophic write-speed drops during intensive operations. In a rebuild scenario, an SMR drive can drop from 150MB/s to 2MB/s, causing the array controller to assume the drive is dead and drop it entirely.

Always ensure you purchase Conventional Magnetic Recording (CMR) drives like Seagate IronWolf (Pro) or WD Red Plus/Pro. Standard "WD Red" (non-plus) drives are often SMR.

📖 Read the full CMR vs SMR guide →

HBA Cards & "IT Mode"

If your motherboard lacks enough SATA ports, do not buy a cheap PCIe SATA expansion card from a non-name brand—they use port multipliers that cause extreme bottlenecks.

Instead, buy a used enterprise Host Bus Adapter (HBA) on eBay (like an LSI Broadcom 9300-8i). You must ensure it is flashed to "IT Mode" (Initiator Target Mode). This disables the card's native hardware raid features and passes the disks directly to the OS, which is absolutely required for software RAID like ZFS.

Storage Enclosures (DAS vs NAS vs Disk Shelves)

DAS (Direct Attached Storage): A \"dummy\" box that holds drives and provides a power supply, connecting to a host machine via USB-C or Thunderbolt. Great for 4 to 8 drive arrays attached to micro-PCs (like an Intel NUC or Mac Mini).

NAS (Network Attached Storage): A dedicated system (like a Synology or TrueNAS box) that handles the RAID computation, filesystem administration, and serves files over network protocols (SMB/NFS).

SAS Disk Shelf (Enterprise): Massive 12-24 bay rackmount enclosures. Instead of connecting to a motherboard, they connect via external SAS cables (SFF-8088 or SFF-8644) to an HBA card in your main server. Loud, power-hungry, but incredibly cheap on the used enterprise market.

📊 Performance Estimate Methodology

Our performance estimates use the Arcserve write-penalty framework, the industry-standard model for comparing RAID throughput. Write speeds are calculated as N / penalty where the penalty reflects parity overhead:

RAID 5: penalty 4×
RAID 6: penalty 6×
ZFS Z1: penalty 3×
ZFS Z2: penalty 5×
SHR: penalty 4.5×
SHR-2: penalty 6.5×
RAID 10: penalty 2×
RAID 0/JBOD: penalty 1×

These are best-case sequential throughput estimates at 250 MB/s per drive. Real-world performance depends on CPU, RAM cache, filesystem overhead, and workload patterns. ZFS's Copy-on-Write architecture and variable stripe width reduce the effective penalty compared to traditional RAID. Full methodology:Performance Methodology.

Browse All Guides & Articles