Network Inventory¶

Quick-reference: Tailscale IP ↔ LAN IP ↔ hostname ↔ role ↔ services.

Purpose: Khi debug issue, nhìn IP → biết ngay machine nào, service nào. Source: Mikrotik DHCP leases + Tailscale status + Komodo TOML (2026-05-18)

Network Topology¶

graph TD
    INET["Internet (PPPoE: 1.53.233.206)"]
    INET --> MK["Mikrotik CHR
192.168.100.1 / 10.10.4.1"]
    MK -->|ether2-home| HOME["10.10.4.0/24
VLAN: Home - IoT/Client devices"]
    MK -->|ether3-lab| LAB["192.168.100.0/24
LAN: Servers/Infra"]
    MK -->|wg-server| WG["10.10.10.0/24 (WireGuard)"]
    MK --> PVE
    subgraph PVE["PVE (Proxmox VE)"]
        UNR["unraid (192.168.100.59)
→ 100.68.251.84"]
        DEV["development (192.168.100.31)
→ 100.126.172.96"]
        MKCHR["Mikrotik CHR (192.168.100.1)"]
    end

Infrastructure Servers (LAN: 192.168.100.x)¶

Client Devices (Home VLAN: 10.10.4.x)¶

Tailscale Network¶

Komodo Periphery Ports¶

Key Service Ports (by Server)¶

unraid (192.168.100.59 / 100.68.251.84)¶

development (192.168.100.31 / 100.126.172.96)¶

Service → Server Mapping¶

unraid stacks¶

applications:  immich, gitea, jellyfin, emby, homepage, protonmail, aria2, delphitools, homeassistant, linkding, searxng, vaultwarden, zalo-tg
database:      pigsty, postgres, postgresus, mariadb, redis, opensearch, meilisearch, qdrant, iceberg
development:   bigquery-emulator, agentmemory
llm:           litellm, 9router, odysseus, qmd
monitoring:    grafana, mimir-loki-tempo, otel-collector, uptime-kuma
networking:    adguard-home, caddy, gluetun, cloudflared-unraid, mosquitto
orchestration: airflow
storage:       romm

development stacks¶

applications:  ntfy, crawl4ai, koodo
database:      clickhouse, memgraph
development:   komodo, rosetta
llm:           usage-plan-mcp
monitoring:    promtail-dev, otel-collector
networking:    adguard-replica, adguard-sync, pmacct, vector
storage:       redpanda

Debug Decision Flow¶

graph TD
    S1["1. Issue trên stack X
→ grep 'server = ' komodo/stacks/{category}/{X}.toml
→ biết target server (unraid / development)"]
    S2["2. Mình đang ở máy nào?
→ hostname → development / unraid / ...
→ ip addr | grep 192.168.100
→ tailscale ip -1"]
    S1 --> S2
    S2 --> S3{"3. So sánh current vs target"}
    S3 -->|SAME machine| LOCAL["act LOCAL:
docker logs, docker compose, etc"]
    S3 -->|DIFF machine| REMOTE["act REMOTE:
km CLI, SSH, or Mikrotik API"]

Local Actions (same machine)¶

docker compose -f /opt/appdata/{service}/docker-compose.yml logs -f
docker ps | grep {service}
journalctl -u {service}
# Volume/data check
ls -la /mnt/user/appdata/{service}/     # unraid
ls -la /opt/appdata/{service}/          # development

Remote Actions (different machine)¶

# Komodo CLI (preferred — works from anywhere)
km-logs -S {stack-name}
km server-stats {server-name}
km deploy stack {stack-name}

# SSH (when need shell access)
ssh root@192.168.100.59    # unraid
ssh lucndm@192.168.100.31  # development

# SSH one-liner
ssh root@192.168.100.59 "docker ps | grep {service}"
ssh root@192.168.100.59 "tail -100 /mnt/user/appdata/{service}/logs/*.log"

# Mikrotik (network-level debug)
curl -sk -u "admin:minhluc1" "http://192.168.100.1/rest/ip/dhcp-server/lease"

Quick Reference: Stack → Where to Debug¶

IP → Hostname Quick Lookup¶

192.168.100.1      → mikrotik (router/gateway)
192.168.100.59     → unraid server
192.168.100.31    → development server
192.168.100.10     → caddy (macvlan)
192.168.100.229    → adguard replica (macvlan)
192.168.100.250    → macvlan shim (unraid→caddy)
100.68.251.84    → unraid (tailscale)
100.126.172.96   → development (tailscale)
100.114.144.55   → pve (tailscale)
100.104.232.25   → mikrotik vpn-exit (tailscale)

VLAN Layout¶

Notes¶

• Tailscale MagicDNS: unraid.tail1137c.ts.net, development.tail1137c.ts.net
• All traffic encrypted via WireGuard (Tailscale)
• Unraid nginx binds 80/443 trên Tailscale IP — services dùng port khác
• PVE is the hypervisor hosting: unraid, development, xpenology, mikrotik CHR as VMs
• Public IP: 1.53.233.206 (PPPoE, dynamic)