Storage Strategy¶

Drive layout, backup topology, and data protection across all hosts.

Drive Inventory¶

NAS (ASUS P8H77-I — 4 SATA ports)¶

Device	Model	Size	Mount	Role	Status

Docker VM (VM 101)¶

Disk	Size	Role

| virtio0 | 100GB | Debian OS, Docker volumes, container data |

VPS (Vultr)¶

Disk	Size	Role

| NVMe | 32GB | OS, Headscale DB, container data |

Spare Drives (at home)¶

Drive	Size	Interface	Location

Storage Architecture¶

NAS Filesystem Layout¶

NAS Drives
├── sda (240GB SSD) ─── LVM root (/)
│   ├── /data/docker/          # Docker data-root (images, volumes)
│   ├── /srv/forgejo/          # Forgejo git data
│   └── /data/coolify/         # Coolify PaaS data
│
├── sdb (2TB Purple) ─── /mnt/purple
│   ├── /mnt/purple/frigate/          # Frigate NVR recordings
│   ├── /mnt/purple/backup/restic/    # Restic REST server data
│   └── /mnt/purple/red-recovery/     # 8TB recovery data (temporary, 1.5TB)
│
├── sdc (8TB Red) ─── /mnt/red8 (after recovery)
│   ├── /mnt/red8/media/        # Music, videos, BD rips
│   ├── /mnt/red8/downloads/    # *arr stack downloads
│   ├── /mnt/red8/data/         # Family photos/videos, old backups
│   ├── /mnt/red8/sync/         # Syncthing data
│   └── /mnt/red8/backup/       # Service backups
│
└── sdd (512GB SSD) ─── /mnt/backup-ssd (after recovery)
    ├── /mnt/backup-ssd/family-videos/   # 45GB (irreplaceable)
    └── /mnt/backup-ssd/family-photos/   # 168GB (irreplaceable)

NAS Symlink Layer (`/srv/`)¶

Services access data through /srv/ symlinks, decoupling mount points from service paths:

/srv/frigate      → /mnt/purple/frigate
/srv/media        → /mnt/red8/media
/srv/downloads    → /mnt/red8/downloads
/srv/data         → /mnt/red8/data
/srv/sync         → /mnt/red8/sync
/srv/backup       → /mnt/red8/backup

NFS Exports (NAS → Docker VM)¶

NAS Export	Docker VM Mount	Used By	Mode

Docker VM fstab entries use defaults,_netdev,nofail to handle NAS unavailability gracefully.

Backup Topology¶

3-2-1 Strategy¶

[Docker VM]                    [NAS]                        [Google Drive]
 Live data                      Restic REST                  rclone crypt
 (Docker volumes)               (/mnt/purple/backup/restic)  (gdrive-crypt:homelab/)
     │                               │                            │
     │  restic backup (nightly) ─────┘                            │
     │                                                            │
     │                          offsite-sync (4:30 AM) ───────────┘
     │
     └── Copy 1 (production)    Copy 2 (local backup)        Copy 3 (offsite)

Backup Schedule (PYT — America/Asuncion)¶

Time	Service	Container	Restic Repo	What's Backed Up

Retention Policy¶

All Restic sidecars use the same retention:

Period	Keep

| Daily | 7 | | Weekly | 4 | | Monthly | 12 |

Automatic pruning runs after every backup. Weekly integrity checks on Sundays.

Restic REST Server¶

# NAS — /mnt/purple/backup/restic/
restic-rest:
  image: restic/rest-server:0.14.0
  ports: "8000:8000"
  options: --private-repos --prometheus
  auth: htpasswd (forces /username/ prefix per repo)
  limits: 256M RAM, 0.5 CPU

All Docker VM sidecars connect to rest:http://augusto:<pass>@192.168.0.12:8000/augusto/<service>.

Offsite Sync Details¶

offsite-sync (NAS container, 4:30 AM)
├── 1. rsync Headscale backups from VPS (SSH, last 48h only)
├── 2. rclone sync Restic repos → gdrive-crypt:homelab/restic/
├── 3. rclone sync Headscale staging → gdrive-crypt:homelab/headscale/
├── 4. ntfy notification (success → cronova-info, failure → cronova-critical)
└── Bandwidth limit: 5 MB/s (RCLONE_BWLIMIT)

Encryption:rclone crypt wraps Google Drive with AES-256. Restore requiresthree secrets: rclone password, rclone salt, and RESTIC_PASSWORD. All stored in Vaultwarden.

Data Criticality¶

Data	Size	Irreplaceable	Copies	Location(s)

Protection Gaps¶

Gap	Risk	Mitigation

Capacity Planning¶

Current Usage¶

Drive	Total	Used	Free	Pressure

| NAS SSD (sda) | 240GB | ~60GB | ~180GB | Low | | NAS Purple (sdb) | 2TB | 1.7TB | 45GB | Critical | | NAS Red (sdc) | 8TB | — | — | Recovery pending | | NAS SSD (sdd) | 512GB | ~0 | 512GB | Unmounted | | Docker VM | 100GB | ~40GB | ~60GB | Low | | VPS | 32GB | ~10GB | ~22GB | Low |

Post-Recovery Projections¶

After 8TB recovery and Purple cleanup:

Drive	Projected Used	Projected Free	Growth Rate

| Purple (sdb) | ~170GB | ~1.6TB | ~5GB/week (Frigate) | | Red (sdc) | ~1.5TB | ~5.8TB | Slow (media additions) | | Backup SSD (sdd) | ~213GB | ~263GB | Static (critical data only) |

Purple will sustain Frigate for ~6 years at current recording rates after recovery frees 1.5TB.

Growth Triggers¶

Trigger	When	Action

Recovery Procedures¶

Quick Reference¶

Scenario	Recovery Source	Procedure	RTO

Full procedures: disaster-recovery.md

Backup verification: backup-test-procedure.md

8TB recovery plan: 8tb-recovery-plan-2026-03-12.md

Backup Scripts¶

All backup scripts live in docker/shared/backup/:

Script	Used By	Purpose

Verification Schedule¶

Frequency	Check	Method