Storage Architecture in a Homelab: RAID, Disks and Data Integrity

This post focuses on how storage is designed and managed within my homelab, including RAID configuration, disk layout, and data integrity considerations.

Storage is one of the most critical components of any system. Unlike stateless services, data loss is often irreversible, so decisions around storage need to be deliberate.

Storage Overview

The system uses a combination of:

• RAID1 (mirroring) for critical data
• A standalone large-capacity disk for media storage

The goal is to balance redundancy, capacity, and simplicity.

RAID Configuration

RAID1 is used for important data, providing redundancy by mirroring data across two disks.

Disk A ─────┐
            ├── RAID1 (mdadm) → Mounted filesystem
Disk B ─────┘

If one disk fails:

• The system continues operating
• Data remains accessible
• The failed disk can be replaced and rebuilt

This provides protection against single disk failure.

Why RAID1

RAID1 was chosen for:

• Simplicity
• Reliability
• Ease of recovery

More complex RAID levels (RAID5/6) introduce additional complexity and rebuild risks, especially in smaller environments.

For this setup, mirroring provides sufficient protection without unnecessary overhead.

Filesystem and Mounting

The RAID array is formatted with a standard Linux filesystem and mounted persistently.

Key considerations include:

• Consistent mount points
• Stable device identification (UUIDs)
• Avoiding reliance on dynamic device names

This ensures the system remains predictable across reboots.

Standalone Storage

A separate high-capacity disk is used for media storage.

This is intentional:

• Media data is non-critical
• Maximising capacity is more important than redundancy

Separating critical and non-critical data simplifies storage design and reduces cost.

Data Integrity Approach

Data integrity is addressed through:

• RAID redundancy for critical data
• SMART monitoring for disk health
• Regular validation of filesystem consistency

This provides multiple layers of protection beyond simple storage.

Migration and Recovery Experience

During migration to a new system, the RAID array needed to be reassembled and mounted correctly.

This involved:

• Identifying disks using persistent IDs
• Reassembling the mdadm array
• Verifying filesystem integrity
• Remounting data safely

This process reinforced the importance of understanding storage at a low level rather than relying on abstractions.

Issues Encountered

Several challenges came up:

• Device name changes across reboots
• Incorrect mount configurations
• Risk of mounting the wrong disk
• Ensuring data was not overwritten during migration

These required careful validation at each step to avoid data loss.

Key Learnings

• Storage design should prioritise data integrity over convenience
• RAID is not a backup, but provides resilience
• Persistent identifiers (UUIDs) are essential
• Understanding underlying storage systems is critical

What’s Next

The next step is exploring network design and segmentation, including how services are structured across the network.