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.

Storage Architecture & Hardware Guide

Last updated: May 2026

Building your first dedicated storage server can be overwhelming. Selecting between NAS, DAS, and SAN architectures, understanding Host Bus Adapters (HBAs) versus hardware RAID controllers, and learning why cards are flashed to "IT Mode" are critical steps in designing a resilient, high-performance homelab or enterprise storage node.

1. Storage Topologies: NAS vs. DAS vs. SAN

The first design choice is deciding how your storage nodes are connected to your computers and networks:

NAS (Network Attached Storage)

A standalone computer containing drives connected to your Local Area Network (LAN). It shares files over standard protocols like SMB (Windows), NFS (Linux), or AFP/WebDAV.

Examples: Synology DiskStations, TrueNAS servers, DIY Custom NAS rigs.

DAS (Direct Attached Storage)

A storage enclosure connected directly to a single host computer via high-speed cables like USB-C, Thunderbolt, or SAS. The host computer manages the filesystem and file sharing.

Examples: External USB hard drive docks, multi-bay USB RAID enclosures, SAS JBOD shelves.

SAN (Storage Area Network)

A specialized, high-speed network that provides servers with direct, block-level access to storage volumes. Shared over dedicated fiber channels or ethernet protocols like iSCSI.

Examples: Enterprise disk arrays, block storage volumes in cloud platforms.

2. Hardware RAID Controllers vs. HBAs

When connecting multiple drives inside a chassis, you need a controller interface card that plugs into your motherboard's PCIe slots:

Hardware RAID Controllers: These cards contain their own onboard processor (I/O processor), RAM cache, and often a battery backup unit (BBU). The card's hardware handles calculations for parity (RAID 5/6/50/60) and presents a single virtual volume to the operating system.

HBAs (Host Bus Adapters): These are simple pass-through cards. They do not calculate parity or manage arrays. Instead, they act as high-speed SAS-to-SATA adapters, presenting each individual hard drive directly to the operating system's kernel.

What is "IT Mode" and why is it necessary?

Most enterprise-grade SAS controllers (like LSI Broadcom cards) can be flashed with two different types of firmware:

  • IR Mode (Integrated RAID): Enables the card's proprietary hardware RAID capabilities. The operating system has no direct visibility into individual drives or raw sectors.
  • IT Mode (Initiator-Target): Disables the RAID BIOS entirely, flashing the card to function as a pure pass-through HBA. Every drive is presented directly to the operating system as an independent block device.

Crucial Rule: If you are running modern software-defined storage systems like ZFS (TrueNAS, Proxmox) or unRAID, you MUST flash your controller card to IT Mode. ZFS and unRAID rely on direct access to the drives' raw sectors to read S.M.A.R.T. health data, verify data blocks with cryptographic checksums, perform disk self-healing, and manage drive spin-down. Hiding the drives behind an IR-mode hardware controller blocks these capabilities and risks massive data corruption during power loss.

3. Storage Operating Systems Comparison

The choice of operating system dictates how files are cached, how parity is computed, and what drives you can mix:

Operating SystemUnderlying EngineMixed Sizes?Key AdvantagesPrimary Drawbacks
TrueNAS (Scale / Core)OpenZFSNo (wastes size diffs)Self-healing data integrity (checksums), high-speed RAM caching (ARC), enterprise stability.Inflexible expansion: Adding capacity requires adding an entire drive group (vdev) of matching sizes. High RAM demand.
unRAIDProprietary JBOD + ParityYes (any sizes)Extremely flexible: mix drive sizes freely. Drives spin down to save power. If multiple disks fail, you only lose files on those specific disks.No striping speed boost: Reads and writes are limited to a single drive's speed. Slow parity updates without a cache SSD. Closed-source.
Synology DSM (SHR)Linux mdraid + LVMYes (auto-slices)Easy mixed-size drive planning, automatic slicing, user-friendly DSM ecosystem.Locked into Synology hardware. Proprietary RAID abstraction cannot easily be recovered on a PC.
OpenMediaVault (OMV)Linux mdadm / Btrfs / ZFSDepends on setupExtremely lightweight, Debian-based, open-source. Runs on low-powered machines (like Raspberry Pi).Requires manual configuration and intermediate Linux sysadmin skills compared to DSM or unRAID.

4. Designing Your First Array: Quick Recommendations

  • For Homelabs & Media Servers (Mixed Sizes): Use unRAID or a **Synology NAS** with SHR. They let you buy whatever drives are on sale and expand capacity one drive at a time.
  • For High-Performance & File Safety (Uniform Sizes): Use TrueNAS Scale (ZFS). Build an array with matching drives (e.g. 6 × 8TB drives in RAID-Z2). You get self-healing data protection and near-instant snapshots.
  • Remember the Parity Rule: In both unRAID and Synology SHR, the size of your parity/redundancy drive must be equal to or larger than your largest data drive. If you buy a new 16TB drive to add to an array of 8TB drives, the 16TB drive must serve as the parity drive!