homelab-optimized/docs/infrastructure/storage.md

💾 Storage Systems

🟡 Intermediate Guide

This document covers the storage architecture, RAID configurations, backup strategies, and data management practices for the homelab infrastructure.

---

🏗️ Storage Architecture Overview

┌─────────────────────────────────────────────────────────────────────────────┐
│                         STORAGE INFRASTRUCTURE                               │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                              │
│  PRIMARY STORAGE                    BACKUP TARGETS                           │
│  ┌─────────────────────┐           ┌─────────────────────┐                  │
│  │     ATLANTIS        │           │      CALYPSO        │                  │
│  │   Synology NAS      │ ──────►   │   Synology NAS      │                  │
│  │                     │  Hyper    │                     │                  │
│  │   8x 16TB RAID 6    │  Backup   │   2x 12TB RAID 1    │                  │
│  │   ≈96TB usable      │           │   ≈12TB usable      │                  │
│  │                     │           │                     │                  │
│  │   + 2x 480GB NVMe   │           │   + 2x 480GB NVMe   │                  │
│  │   (SSD Cache)       │           │   (SSD Cache)       │                  │
│  └─────────────────────┘           └─────────────────────┘                  │
│            │                                  │                              │
│            │                                  │                              │
│            ▼                                  ▼                              │
│  ┌─────────────────────────────────────────────────────────┐                │
│  │                    BACKBLAZE B2                          │                │
│  │              Cloud Offsite Backup                        │                │
│  │           Encrypted, Versioned Storage                   │                │
│  └─────────────────────────────────────────────────────────┘                │
│                                                                              │
│  SECONDARY STORAGE                                                          │
│  ┌───────────────┐  ┌───────────────┐  ┌───────────────┐                   │
│  │    GUAVA      │  │   SETILLO     │  │   PROXMOX    │                   │
│  │  RAID 1 HDD   │  │  Single 1TB   │  │   Local SSD  │                   │
│  │  + NVMe SSD   │  │               │  │              │                   │
│  └───────────────┘  └───────────────┘  └───────────────┘                   │
│                                                                              │
└─────────────────────────────────────────────────────────────────────────────┘

---

📊 Storage Summary

Host	Total Raw	Usable	RAID Level	Purpose
Atlantis	128TB (8x16TB)	~96TB	RAID 6	Primary storage, media
Calypso	24TB (2x12TB)	~12TB	RAID 1	Backup, development
Guava	6TB+	~3TB	RAID 1	AI/ML, compute
Setillo	1TB	1TB	Single	Monitoring
Proxmox	~500GB	500GB	Local SSD	VM storage

---

🏛️ Atlantis - Primary Storage

Hardware Configuration

Component	Specification
NAS Model	Synology DS1823xs+
Drive Bays	8x 3.5" hot-swap
Drives	8x Seagate IronWolf Pro 16TB (ST16000NT001)
Cache	2x WD Black SN750 480GB NVMe
RAID Level	RAID 6 (dual parity)
Raw Capacity	128TB
Usable Capacity	~96TB
Fault Tolerance	2 drive failures

RAID 6 Benefits

RAID 6 Configuration:
┌────┬────┬────┬────┬────┬────┬────┬────┐
│ D1 │ D2 │ D3 │ D4 │ D5 │ D6 │ P1 │ P2 │  ← Data + Dual Parity
├────┼────┼────┼────┼────┼────┼────┼────┤
│ D1 │ D2 │ D3 │ D4 │ D5 │ P1 │ P2 │ D6 │  ← Parity distributed
├────┼────┼────┼────┼────┼────┼────┼────┤
│ D1 │ D2 │ D3 │ D4 │ P1 │ P2 │ D5 │ D6 │
└────┴────┴────┴────┴────┴────┴────┴────┘

✅ Survives 2 simultaneous drive failures
✅ Good read performance
✅ 6 drives worth of usable space (75% efficiency)
⚠️ Slower writes due to parity calculation

Volume Layout

/volume1/ (Atlantis - ~96TB usable)
│
├── /docker/              # Container persistent data
│   ├── plex/
│   ├── immich/
│   ├── grafana/
│   └── ... (all stack data)
│
├── /media/               # Media library
│   ├── movies/           # 4K + 1080p movies
│   ├── tv/               # TV series
│   ├── music/            # Music library
│   └── audiobooks/       # Audiobook collection
│
├── /photos/              # Immich photo library
│   ├── library/          # Organized photos
│   └── upload/           # Incoming uploads
│
├── /documents/           # Paperless-NGX
│   ├── consume/          # Incoming documents
│   └── archive/          # Processed documents
│
├── /backups/             # Local backup storage
│   ├── calypso/          # Cross-NAS backups
│   └── vm-snapshots/     # VM backup images
│
└── /archive/             # Long-term cold storage
    └── old-projects/

NVMe SSD Cache

• Type: Read-write cache
• Drives: 2x WD Black SN750 480GB
• Configuration: RAID 1 (mirrored for safety)
• Purpose: Accelerate frequently accessed data

---

🏢 Calypso - Secondary Storage

Hardware Configuration

Component	Specification
NAS Model	Synology DS723+
Drive Bays	2x 3.5" hot-swap
Drives	2x Seagate IronWolf Pro 12TB (ST12000NT001)
Cache	2x WD Black SN750 480GB NVMe
RAID Level	RAID 1 (mirrored)
Raw Capacity	24TB
Usable Capacity	~12TB
Fault Tolerance	1 drive failure

RAID 1 Benefits

RAID 1 Configuration:
┌────────────────┐  ┌────────────────┐
│    Drive 1     │  │    Drive 2     │
│    (12TB)      │◄─►    (12TB)      │  ← Mirror
│                │  │                │
│  All data is   │  │  Exact copy    │
│  written to    │  │  of Drive 1    │
│  both drives   │  │                │
└────────────────┘  └────────────────┘

✅ Survives 1 drive failure
✅ Fast read performance (can read from either)
✅ Simple recovery (just replace failed drive)
⚠️ 50% storage efficiency

Volume Layout

/volume1/ (Calypso - ~12TB usable)
│
├── /docker/              # Container persistent data
│   ├── gitea/
│   ├── firefly/
│   ├── arr-suite/
│   └── ... (dev stacks)
│
├── /apt-cache/           # APT-Cacher-NG
│   └── cache/            # Debian package cache
│
├── /backups/             # Backup destination
│   ├── atlantis/         # Hyper Backup from Atlantis
│   └── databases/        # Database dumps
│
└── /development/         # Development data
    ├── repos/            # Git repositories
    └── projects/         # Project files

---

🖥️ Other Storage Systems

Guava - AI/ML Workstation

Component	Specification
Primary	1TB NVMe SSD (OS + fast storage)
Secondary	2x HDD in RAID 1 (~3TB usable)
Purpose	AI model storage, datasets, compute scratch

Setillo - Monitoring

Component	Specification
Storage	1TB single drive
Purpose	Prometheus metrics, AdGuard data
Note	Non-critical data, can be rebuilt

Proxmox - VM Host

Component	Specification
Storage	~500GB local SSD
Purpose	VM disk images
Backup	VMs backed up to Atlantis

---

📦 Backup Strategy

3-2-1 Rule Implementation

Rule	Implementation	Status
3 Copies	Original + Calypso + Backblaze	✅
2 Media Types	NAS HDDs + Cloud	✅
1 Offsite	Backblaze B2	✅

Backup Flow

┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│  ATLANTIS   │────►│   CALYPSO   │────►│ BACKBLAZE   │
│  (Primary)  │     │  (Local)    │     │    B2       │
│             │     │             │     │  (Offsite)  │
│  Original   │     │  Hyper      │     │  Cloud      │
│  Data       │     │  Backup     │     │  Backup     │
└─────────────┘     └─────────────┘     └─────────────┘
     │                    │                    │
     │                    │                    │
     ▼                    ▼                    ▼
  Immediate           < 24 hours          < 24 hours
   Access              Recovery            Recovery

Backup Software

Tool	Source	Destination	Schedule
Synology Hyper Backup	Atlantis	Calypso	Daily
Synology Cloud Sync	Atlantis	Backblaze B2	Daily
Synology Hyper Backup	Calypso	Backblaze B2	Weekly

What Gets Backed Up

Data Type	Priority	Frequency	Retention
Docker configs	Critical	Daily	30 days
Databases	Critical	Daily	30 days
Photos (Immich)	High	Daily	Forever
Documents	High	Daily	1 year
Media library	Medium	Weekly	Latest only
VM snapshots	Medium	Weekly	4 versions
Logs	Low	Not backed up	N/A

Recovery Time Objectives

Scenario	RTO Target	Recovery Method
Single file recovery	< 1 hour	Hyper Backup restore
Service recovery	< 4 hours	Docker volume restore
Full NAS recovery	< 24 hours	Bare metal + B2 restore
Disaster recovery	< 48 hours	New hardware + B2 restore

---

📂 Shared Storage (NFS/SMB)

Network Shares

Share	Protocol	Host	Access	Purpose
/media	SMB	Atlantis	Read-only (most), RW (arr)	Media streaming
/photos	SMB	Atlantis	RW (Immich user)	Photo backup
/docker	NFS	Atlantis	RW (Docker hosts)	Container data
/backups	SMB	Calypso	RW (backup service)	Backup destination

Docker Volume Mounts

Containers access NAS storage via NFS mounts:

# Example: Plex accessing media
volumes:
  - /volume1/docker/plex:/config
  - /volume1/media:/media:ro

Permission Model

NAS User: docker (UID 1000)
├── Owns /volume1/docker/
├── Read access to /volume1/media/
└── Write access to specific paths

NAS User: media (UID 1001)
├── Write access to /volume1/media/
└── Used by *arr suite for downloads

---

📈 Storage Monitoring

Metrics Collected

Metric	Tool	Alert Threshold
Disk usage	Prometheus + Node Exporter	> 85%
RAID health	Synology DSM	Degraded
Drive SMART	Synology DSM	Warning/Critical
I/O latency	Prometheus	> 100ms
Backup status	Hyper Backup	Failed

Grafana Dashboard

Storage dashboard shows:

• Volume utilization trends
• I/O throughput
• RAID rebuild status
• Drive temperatures
• Backup completion status

---

🔮 Storage Expansion Plan

Current Utilization

Host	Used	Total	% Used
Atlantis	~60TB	96TB	62%
Calypso	~12TB	12TB	~100%

Future Expansion Options

1. Atlantis: Already at max capacity (8 bays)

   • Replace 16TB drives with larger (24TB+) when available
   • Add expansion unit (DX517)

2. Calypso: At capacity

   • Replace 12TB drives with 20TB+ drives
   • Consider migration to larger NAS

3. New NAS: For cold/archive storage

   • Lower-powered unit for infrequent access
   • RAID 5 acceptable for archive data

---

🛠️ Maintenance Tasks

Regular Maintenance

Task	Frequency	Procedure
SMART check	Weekly	Review DSM health
Scrub	Monthly	Synology scheduled task
Backup verification	Monthly	Test restore of random files
Capacity review	Quarterly	Plan for growth

Drive Replacement Procedure

1. Identify failed drive via DSM notification
2. Order replacement (same or larger capacity)
3. Hot-swap failed drive
4. Monitor rebuild (can take 24-48 hours for large arrays)
5. Verify RAID health after rebuild completes

---

📚 Related Documentation

• Host Infrastructure: Server specifications
• Security Model: Backup encryption details
• Network Architecture: NFS/SMB networking

---

Storage infrastructure is critical. Regular monitoring and proactive maintenance prevent data loss.