# KaburuLINK — Unraid NAS

Local Unraid box. Previously a full NAS/Docker host. Now in wind-down state — services
removed except Immich and some old file shares pending review and deletion.

## Specs & Access

| Item | Detail |
|------|--------|
| Hostname | KaburuLINK |
| Model | Lincsstation N1 (Techvision TVI8310X, Intel Celeron N5105 @ 2GHz, 16GB LPDDR4) |
| IP | 192.168.0.250 (LAN only — not on Tailscale) |
| OS | Unraid 7.3.0 Plus |
| SSH | root@192.168.0.250 via truenas_id_ed25519 |
| SNMP | Community: kaburu, trap sink: 192.168.0.253 |
| Boot | USB flash (mmcblk0, 115GB) |

## Storage Layout

### Parity

| Device | Model | Serial | Role | Health |
|--------|-------|--------|------|--------|
| sdb | Fanxiang S101 2TB 2.5" SATA SSD | MX_00000000000030180 | **Parity** | ⚠️ SMART error — 14 reallocated sectors, max temp 54°C (thermal damage) |

### Array Disks (healthy)

| Device | Model | Serial | Health |
|--------|-------|--------|--------|
| nvme1n1 | Crucial CT2000P3PSSD8 2TB | 2308E6B2F007 | ✅ 6% wear, 35°C |
| nvme0n1 | Crucial CT2000P3PSSD8 2TB | 2308E6B2F0B7 | ✅ 4% wear, 35°C |
| nvme2n1 | Crucial CT2000P3PSSD8 2TB | 2308E6B2F105 | ✅ 2% wear, 36°C |
| nvme3n1 | Crucial CT2000P3PSSD8 2TB | 2318E6D29A98 | ✅ 0% wear, 34°C |

Array filesystem: ZFS. Default fsType per disk.cfg.

### Cache

| Device | Model | Serial | Role | Health |
|--------|-------|--------|------|--------|
| sdc | Fanxiang S101 2TB 2.5" SATA SSD | MX_00000000000090048 | **Cache (ZFS pool: cache, sdc1)** | ⚠️ 14 reallocated sectors, max temp 54°C — SMART PASSED but thermally damaged |

Cache is nearly empty — 128K used of 1.8TB (1%). Mounted at /mnt/cache.
Cache ZFS pool last scrubbed 2026-04-01, 0 errors.

### Unassigned / Boot

| Device | Notes |
|--------|-------|
| sda | USB Flash Disk — Unraid boot device |
| mmcblk0 | Internal eMMC 115GB — likely OS/boot media |

**Note:** sdc (cache) appears as "Dev 1 — unassigned" in the Unraid array panel because it
is managed as a separate cache pool, not part of the main array. This is normal Unraid behaviour.

---

## Drive Health Summary (SMART — 2026-05-25)

| Device | Role | SMART | Reallocated | Max Temp | Notes |
|--------|------|-------|-------------|----------|-------|
| sdb | Parity | ⚠️ ERROR | 14 | 54°C | Thermally damaged, still functioning |
| sdc | Cache | PASSED | 14 | 54°C | Thermally damaged, SMART OK for now |
| nvme0-3 | Array | ✅ PASSED | 0 | 36°C | All healthy |

Both Fanxiang SSDs were cooked in a rack with poor airflow (~March 2026).
Replacement deferred — SSD prices high. Neither has failed outright.

**Risk:** If parity (sdb) fails, array loses redundancy immediately. NVMe array data intact
but unprotected until parity is rebuilt on a replacement drive.

### Replacement plan
- Need 2× 2.5" SATA SSDs (2TB minimum) — direct replacements for the Fanxiang S101s
- One for parity (sdb slot), one for cache (sdc slot)
- Layout confirmed: NVMe drives were always data array (disk1-4), never cache
- When replacing: use Unraid Tools → New Config or swap and run parity rebuild
- Deferred until SSD prices improve

---

## Why Hermes Flags Parity Alerts

Hermes monitors Unraid via SSH (unraid.py connector). It reports the SMART error on the
parity drive. **This is known and expected** — do not escalate urgently. Report status only
until replacement is planned and actioned.

---

## Active Services

| Service | Status | Notes |
|---------|--------|-------|
| Immich | ✅ Running | Docker — immich + PostgreSQL_Immich containers |
| ClamAV | ⛔ Stopped | Not in use |
| File shares | ⚠️ Pending audit | See below |

### Immich

Photos library (~70GB) stored on Unraid array (NVMe — healthy).
Postgres DB in appdata on cache (sdc).
Nightly backup cron at 03:00:
- DB dump → `/mnt/tank/unraid-backup/immich-backup/postgres/` (7 days retention)
- Appdata rsync → `/mnt/tank/unraid-backup/immich-backup/appdata/`

**TODO:** Migrate Immich to permanent home — TrueNAS or Z840. No decision made yet.

---

## File Shares — Pending Audit

23 shares total (5 public SMB, 3 public NFS per dashboard). Needs full review before
box is decommissioned. Known shares under `/mnt/user/`:

| Share | Security | Notes |
|-------|----------|-------|
| 3D Printing | — | |
| DATA | Public | |
| Documents | Secure | |
| Images | — | |
| Music | Public | |
| Photos | — | Immich library |
| Renee-Server | Secure | Renee's network share |
| SHARE | — | |
| Web Stuff | Public | Old web files — likely deletable |
| appdata | Public | Docker appdata |
| backup_folder | Secure | Backup from QNAS / Google Drive |
| cloud-stuff | Secure | |
| clouddrive | — | |
| docker-cache | — | |
| dockershit | Secure | |
| domains | Secure | Saved VM instances |
| downloads | Secure | Downloaded file dump |
| frigate | — | |
| isos | — | ISO images |
| mega | — | Mega sync files |
| old-nas-stuff | Secure | |
| system | Secure | System data |
| vm-stuff | Public | |

**Action:** Audit each share, confirm deletable, remove. Then migrate Immich. Then decommission or repurpose box.

---

## Decommission Checklist

- [ ] Audit and delete file shares
- [ ] Migrate Immich to TrueNAS or Z840
- [ ] Remove Unraid from Hermes monitoring (or repurpose box)
- [ ] Add to Tailscale if keeping long-term