Skip to main content

​
Summary

This page documents each node in the MOOD MNKY Proxmox cluster from a hardware and capability perspective. Source of truth:
  • Proxmox API snapshots summarized in proxmox-terraform/CLUSTER-NODES-HARDWARE.md.
  • Per-node hardware snapshots under /root/hardware-snapshots/<node>/<timestamp>/ on each host (collected for CODE-MNKY, CASA-MNKY, DATA-MNKY, and STUD-MNKY; PRO-MNKY pending while offline).
  • LXC and service mappings from the homelab “as code” docs and runbooks.

​
Node table

NodeStatusCPU modelCores / threadsRAM (approx)Root storage (approx)Primary local storage (active)
CODE-MNKYOnlineAMD Ryzen 5 4600G (Radeon)6 / 12125 GiB~457 GiBCODE-MAIN-zfs, CODE-BKP-zfs, local, local-zfs, hyper-mnky-shared (NFS)
CASA-MNKYOnlineAMD Ryzen 5 4600G (Radeon)6 / 1262 GiB~1.11 TiBlocal-zfs, local, hyper-mnky-shared (NFS)
DATA-MNKYOnlineAMD Ryzen 7 5700X8 / 1662.7 GiB~1.72 TiBlocal-zfs, local, hyper-mnky-shared (NFS)
STUD-MNKYOnlineAMD Ryzen 7 5700G (Radeon)8 / 1662 GiB~445 GiBSTUD-zfs, local-zfs, local, hyper-mnky-shared (NFS)
PRO-MNKYOffline(captured when online)———(to be filled when node is online)
All online nodes run Proxmox VE 8.4.17 on kernel 6.8.12-19-pve, with EFI boot and Secure Boot off.

​
CODE-MNKY

CODE-MNKY is the primary GPU and stack host and the main focus for deep introspection.

​
Role

  • Hosts the main LXC stack (VMIDs 3099–3104) including:
    • GPU dev LXC (3099) with Codex CLI and dev tooling.
    • Automation stack (3100) with AWX and Semaphore.
    • AI stack (3101) with Ollama, Flowise, n8n, MinIO, and self-hosted Supabase.
    • Gaming and Sunshine stack (3102).
    • Media stack (3103) with Jellyfin, *arr suite, and qBittorrent.
    • Dedicated Supabase stack (3104) backing AI workloads as needed.
  • Target node for most Terraform-created LXCs (target_node = "CODE-MNKY").

​
Hardware profile (snapshot-based)

Based on /root/hardware-snapshots/CODE-MNKY/20260316T152000Z/:
  • CPU:
    • AMD Ryzen 5 4600G, 6 cores / 12 threads, single socket.
    • AVX2, AES-NI, AMD-V virtualization, extensive CPU mitigations.
  • Memory:
    • ~125 GiB installed; single NUMA node (node0).
    • DIMM layout, speeds, and vendors in dmidecode-memory.txt.
  • Storage:
    • CODE-MAIN-zfs (~3.6 TiB, NVMe-backed) – main data pool for LXCs/VMs.
    • CODE-BKP-zfs (~464 GiB, HDD-backed) – backup/secondary pool.
    • rpool (~472 GiB, SSD-backed) – root filesystem.
  • GPU:
    • NVIDIA Tesla P40 (Pascal), driver 560.35.03, CUDA 12.6.
    • Full details in nvidia-smi-q.txt.
  • Network:
    • Multiple physical NICs and bridges, with offload and link details in ethtool-summary.txt.
    • Routes and addressing in ip-addr.txt and ip-route.txt.
  • PCI/USB:
    • Full bus layout in lspci-nnvv.txt, lspci-tree.txt, lsusb.txt, lsusb-tree.txt.
For a full narrative deep dive, see /infra/data-center/code-mnky-node.

​
CASA-MNKY

CASA-MNKY provides additional general-purpose capacity and shared storage access.

​
Role

  • General-purpose compute and storage node.
  • Participates fully in Proxmox quorum and shared NFS (hyper-mnky-shared).
  • Suitable for migrating LXCs and VMs off CODE-MNKY if needed.

​
Hardware profile (snapshot-based)

Based on /root/hardware-snapshots/CASA-MNKY/20260316T152857Z/ (plus cluster summary):
  • CPU: AMD Ryzen 5 4600G (6 cores, 12 threads).
  • Memory: ~62 GiB.
  • Root filesystem: ~1.11 TiB.
  • Storage (active):
    • local-zfs (zfspool).
    • local (dir).
    • hyper-mnky-shared (NFS).
For a full narrative deep dive, see /infra/data-center/casa-mnky-node.

​
DATA-MNKY

DATA-MNKY is optimized for core count and disk capacity.

​
Role

  • High-core-count node suited for CPU-bound workloads.
  • Larger root disk (~1.72 TiB) for data-heavy tasks.

​
Hardware profile (snapshot-based)

Based on /root/hardware-snapshots/DATA-MNKY/20260316T152848Z/ (plus cluster summary):
  • CPU: AMD Ryzen 7 5700X (8 cores, 16 threads), 1 socket, AMD-V.
  • Memory: ~62.7 GiB.
  • Root pool: rpool (~1.81 TiB) mirrored over two NVMe devices.
  • Storage (active): local ZFS + shared NFS (hyper-mnky-shared).
For a full narrative deep dive, see /infra/data-center/data-mnky-node.

​
STUD-MNKY

STUD-MNKY provides additional compute and its own dedicated ZFS pool.

​
Role

  • Additional compute node with a dedicated ZFS pool (STUD-zfs).
  • Useful for isolating particular stacks or experiments.

​
Hardware profile (snapshot-based)

Based on /root/hardware-snapshots/STUD-MNKY/20260316T152852Z/ (plus cluster summary):
  • CPU: AMD Ryzen 7 5700G (8 cores, 16 threads), 1 socket, AMD-V.
  • Memory: ~62 GiB.
  • Pools:
    • STUD-zfs (~3.62 TiB) – dedicated data pool.
    • rpool (~460 GiB) – root pool.
  • Storage (active): local ZFS + shared NFS (hyper-mnky-shared).
For a full narrative deep dive, see /infra/data-center/stud-mnky-node.

​
PRO-MNKY

PRO-MNKY is currently offline from the perspective of detailed hardware documentation.

​
Role

  • Additional node in the cluster; hardware details to be captured once the node is brought online.

​
Next steps

When PRO-MNKY is online: 
sudo /root/proxmox-ansible/scripts/collect-node-hardware.sh
This will produce: 
/root/hardware-snapshots/PRO-MNKY/<timestamp>/
You can then mirror the structure above (identity, CPU, memory, storage, network, GPU if any, PCI/USB) to complete PRO-MNKY’s profile.

​
Keeping node docs current

When hardware changes on any node:
  1. Run the collector on that node.
  2. Update the summary row in the table at the top of this page.
  3. Refresh the relevant per-node section with details from the latest snapshot.
  4. If CODE-MNKY changed significantly, also revisit /infra/data-center/code-mnky-node.