Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.moodmnky.com/llms.txt

Use this file to discover all available pages before exploring further.

This page records a live inventory (via Proxmox on CODE-MNKY) of the three LXCs called out for review: 300 (mnky-automation-stack), 301 (mnky-ai-stack), and 3001 (pegaprox). Use it with the Data Center Map and CODE-MNKY node pages.

​
Executive summary

LXCNameRuntime stateFinding
300mnky-automation-stackEmpty — no Docker/containerd; only base OS (SSH, systemd, postfix)Intended AWX + Semaphore not installed/running. Tags and description are aspirational or left after migration.
301mnky-ai-stackEmpty — no Docker; ~381 MiB used on 2 TiB rootfsIntended Ollama, Flowise, n8n, Supabase, MinIO not present. Supabase and n8n run on QEMU VMs 3055/3056 on the same node instead — see redundancy section.
3001pegaproxActive — Python PegaProx v0.7.0 (Proxmox cluster UI / management)Listens on 5000, 5001, 5002; user pegaprox. Not redundant with Traefik/Coolify; different purpose (Proxmox helper UI).
Media: None of these LXCs run Jellyfin, *arr, or torrent clients. Production media is on MOOD-MNKY LXC 120 (Media stack).

​
Proxmox resource summary

From pct config (CODE VLAN tag=30, bridge vmbr0):

​
LXC 300 — mnky-automation-stack

  • Resources: 4 vCPU, 8 GiB RAM, 128 GiB rootfs on CODE-MAIN-zfs
  • Features: nesting, keyctl, fuse
  • Tags: awx, docker, mnky-automation-stack, semaphore, terraform
  • Live services: No application stack; docker.service inactive; /opt empty

​
LXC 301 — mnky-ai-stack

  • Resources: 8 vCPU, 32 GiB RAM, 2000 GiB rootfs on CODE-MAIN-zfs
  • Features: nesting, keyctl, fuse
  • Tags: docker, flowise, minio, mnky-ai-stack, n8n, ollama, supabase, terraform
  • GPU: Not visible in pct config at inventory time (no lxc.cgroup2.devices / passthrough lines); description text mentions P40 — verify before assuming GPU workloads
  • Live services: No Docker stack; disk almost entirely unused

​
LXC 3001 — pegaprox

  • Resources: 4 vCPU, 4 GiB RAM, 16 GiB rootfs, unprivileged, nesting=1
  • Processes: /opt/PegaProx/venv/bin/python … pegaprox_multi_cluster.py and pegaprox/api/.ssh_ws_server.py
  • Ports: 5000, 5001, 5002 (TCP), SSH 22

​
Redundancy vs current architecture

CapabilityOld intent (tags / mnky-docs 31xx)Current live placement
SupabaseTagged on LXC 301VM 3055 mnky-supabase-prod (QEMU on CODE-MNKY)
n8nTagged on LXC 301VM 3056 mnky-n8n-prod (QEMU on CODE-MNKY)
Ollama / Flowise / MinIOTagged on LXC 301Not running in LXC 301 at inventory time — confirm Coolify/other hosts or decommissioned
Automation (AWX / Semaphore)LXC 300Not running in LXC 300
MediaHistorically “media stack” in old doc VMIDsMOOD-MNKY LXC 120 mnky-media-stack
Documentation drift: Older mnky-docs tables used VMIDs 3099–3104. Live CODE-MNKY currently exposes LXCs 300, 301, 3001 and VMs 3055, 3056 only — Terraform/docs should be reconciled or marked “manual drift.”

​
Same-node redesign options (CODE-MNKY only)

These are options, not mandates:
  1. Decommission or repurpose LXC 300/301
    If automation and AI stacks are permanently elsewhere, snapshot, then destroy or shrink rootfs to reclaim CODE-MAIN-zfs and RAM. Alternatively re-run Ansible to actually install AWX/Semaphore (300) and/or GPU Ollama+Flowise (301) if you still want them on this node.
  2. Keep 301 as dedicated GPU host
    If you bring back Ollama on-node, provision one clear stack (compose + GPU passthrough in pct config), and do not duplicate Supabase/n8n (keep 3055/3056 canonical).
  3. Keep PegaProx (3001)
    Continue if operators use it for Proxmox cluster management; document URLs and firewall. If replaced by portal/another tool, plan migration before shutdown.
  4. Granular LXCs on the same node
    Split into e.g. lxc-ollama-gpu, lxc-flowise, lxc-automation — each with smaller disks and pinned compose — after you decide what must stay on CODE-MNKY vs VMs.

​
Media teardown (300 / 301 / 3001)

No action required on these three LXCs for media: no media containers or images were found. Media removal work belongs under MOOD-MNKY LXC 120 and legacy references only.

​
Verification commands (operators)

On CODE-MNKY as root: 
pct config 300 && pct config 301 && pct config 3001
pct exec 300 -- systemctl status docker --no-pager
pct exec 301 -- systemctl status docker --no-pager
pct exec 3001 -- ss -tulpn
qm list   # VMs 3055, 3056

​
Data source

Inventory date: 2026-03-27. Source: live pct / pct exec / ps on CODE-MNKY (10.3.0.10). Re-run after any provisioning change.