# Storage Drives

Live-verified 2026-04-18. Every drive currently in the homelab, grouped by host. For RAID / volume layout and backup flows see [atlantis-storage.md](atlantis-storage.md) and [storage-topology](../diagrams/storage-topology.md).

## Atlantis (DS1823xs+)

**Bulk storage — Volume 1 (Btrfs, RAID 6, encrypted)**

| Slot | Model | Capacity | Role |
|---|---|---|---|
| 1–8 | Seagate IronWolf Pro (SATA) | 8× 16 TB | 128 TB raw / 84 TB usable |

**NVMe — Volume 2 (Btrfs, RAID 1)**

| Model | Capacity | Slot | Purpose |
|---|---|---|---|
| 2× Crucial CT1000P310SSD8 | 2× 1 TB | M.2 adapter card | Docker configs, databases, metadata (885 GB usable) |

**NVMe — Volume 3 (Btrfs, RAID 1)**

| Model | Capacity | Slot | Purpose |
|---|---|---|---|
| 2× Synology SNV5420-400G | 2× 400 GB | M.2 built-in | SABnzbd working directory, high-I/O temp (348 GB usable) |

## Calypso (DS723+)

**Volume 1 (Btrfs, RAID 1, encrypted)**

| Slot | Model | Capacity | Role |
|---|---|---|---|
| 1–2 | Seagate IronWolf Pro (SATA) | 2× 12 TB | 24 TB raw / ~11 TB usable |

**NVMe cache (RAID 1)**

| Model | Capacity | Purpose |
|---|---|---|
| 2× Crucial P3 Plus | 2× 500 GB | Read/write cache for Volume 1 |

## Setillo (DS223j)

**Volume 1 (Btrfs, RAID 1)**

| Slot | Model | Capacity | Role |
|---|---|---|---|
| 1–2 | WD Gold | 2× 10 TB | 20 TB raw / ~8.8 TB usable — off-site backup target |

## Guava (TrueNAS SCALE)

**boot-pool (ZFS, single device)**

| Model | Capacity | Purpose |
|---|---|---|
| WD Black SN770 NVMe | 464 GB (433 GB free) | OS / boot |

**data pool (ZFS mirror, dedup 1.67×)**

| Model | Capacity | Purpose |
|---|---|---|
| 2× WD Blue SA510 SATA SSD | 2× 4 TB | User data (3.62 TB raw, 2.16 TB used, 1.47 TB free) |

Datasets under `data/`: `guava_turquoise` (3 TB personal data), `photos`, `jellyfin`, `llama` (LLM model cache), `plane-data`, `cocalc`, `fasten`, `openproject`, `medical`. NFS mount `atlantis_media` re-exposes Atlantis Volume 1.

## Shinku-Ryuu (workstation)

| Drive | Role |
|---|---|
| NVMe SSD (primary) | Windows OS, games |
| NVMe SSD (secondary) | Game library |
| SATA SSD | Bulk / scratch |

(Exact models not tracked here — managed manually on the Windows host.)

## Concord NUC

| Drive | Model | Capacity |
|---|---|---|
| SATA SSD | Toshiba VX500 | 238 GB |

## Homelab VM

| Drive | Capacity | Role |
|---|---|---|
| Proxmox virtual disk | ~500 GB | OS + container volumes (backed by PVE host ZFS) |

## Matrix-Ubuntu VM

| Drive | Capacity | Role |
|---|---|---|
| LVM LV on Synology virtual disk | 1 TB (~1005 GB) | OS + all container data (Synapse, Mastodon, LiveKit, NPM certs) |

## Pi 5 (Vish)

| Drive | Role |
|---|---|
| NVMe SSD via PCIe HAT | Boot + Uptime Kuma / DIUN data |

## Pi 5 (Jellyfish)

| Drive | Role | Status |
|---|---|---|
| Internal NVMe | Boot | active |
| External LUKS-encrypted SSD | Photo / media | **recovery in progress** (LUKS closed, needs passphrase + `e2fsck`) — see `project_jellyfish_fsck` memory and `/home/lulu/FSCK-RECOVERY-NOTES.md` on the host |

## Olares (K3s laptop)

| Drive | Role |
|---|---|
| Internal NVMe | Ubuntu root + K3s local storage |

NFS mounts from Atlantis/Calypso back most K3s-app persistent volumes (Ollama model cache, Jellyfin metadata, Tdarr cache).

## Seattle VPS

| Drive | Capacity | Role |
|---|---|---|
| Contabo virtual disk | 290 GB (~110 GB free) | OS + all container data |

## Setillo `/opt` (Entware)

Not a drive per se — `/volume1/@entware` is bind-mounted to `/opt` via systemd `opt.mount`. ~260 MB Entware packages, survives DSM upgrades so long as the systemd unit is preserved. See [hosts/setillo.md](../hosts/setillo.md).

## Backups in flight

- Atlantis Volume 1 → Calypso (HyperBackup, weekly/daily depending on dataset)
- Atlantis/Calypso `/volume1/docker` → Setillo via Syncthing (real-time)
- Atlantis photos + media → Backblaze B2 bucket `vk-atlantis` (HyperBackup S3, weekly)
- Calypso docker → Backblaze B2 bucket `vk-concord-1` (HyperBackup S3, daily)

See [infrastructure/backup-strategy.md](../infrastructure/backup-strategy.md) for RTO/RPO and restore procedures.

## SMART / health

- Synology NAS: DSM Storage Manager SMART reports + Scrutiny container on homelab-vm (`scrutiny.vish.gg`).
- Guava: TrueNAS SCALE native SMART; NetData 339 MB dataset pinned to the data pool.
- NUC / Pi / VMs: `smartctl` ad hoc; no centralized collection.

## Replacement / warranty tracking

Actual purchase dates and serial numbers are tracked in NetBox (`nb.vish.gg`) rather than this file. Check DCIM → Devices → the host's "Inventory" tab for per-drive serials.