d-bis/proxmox
4
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity
Files
main
proxmox/docs/11-references/HARDWARE_INVENTORY_MASTER.md
defiQUG e4c9dda0fd
Some checks failed
Deploy to Phoenix / deploy (push) Has been cancelled
Details
chore: update submodule references and documentation 
- Marked submodules ai-mcp-pmm-controller, explorer-monorepo, and smom-dbis-138 as dirty to reflect recent changes.
- Updated documentation to clarify operator script usage, including dotenv loading and task execution instructions.
- Enhanced the README and various index files to provide clearer navigation and task completion guidance.

Made-with: Cursor
2026-03-04 02:03:08 -08:00

3.6 KiB
Raw Permalink Blame History

Hardware inventory — master reference

Last updated: 2026-03-03
Purpose: Single source of truth for total server, workstation, switching, and gateway hardware.

Servers

Type Count Model / notes Role
Dell PowerEdge R630 13 2-socket, ECC, 2.5" bays Proxmox cluster nodes (r630-01 … r630-13); IP plan 192.168.11.11–23. Currently 2 active (r630-01, r630-02); r630-03/04 documented powered off. See R630_13_NODE_DOD_HA_MASTER_PLAN.md.
Dell PowerEdge R750 3 2U, 2-socket, NVMe/SSD, GPU-capable GPU/AI tier — assign IPs 192.168.11.24–.26; 2–4 PCIe GPUs per node (e.g. A6000, L40S) for ML/inference. Alternative: 2× XE9680 (16× A100 80GB) for heavy AI/training instead of 3× R750.

R630 IP plan (13 nodes): 192.168.11.11 – 192.168.11.23 (r630-01 … r630-13).
R750 IP suggestion: 192.168.11.24 – 192.168.11.26 (r750-01, r750-02, r750-03). Document in config/ip-addresses.conf as nodes are brought online.

Workstations

Type Count Model / notes Role
Dell Precision 7920 2 Tower workstation, multi-GPU capable Admin, dev, or build workstations; assign IPs in 192.168.11.x (e.g. .30–.31); connect to management LAN and/or 10G if NICs support.

7920 IP suggestion: 192.168.11.30 – 192.168.11.31 (workstation-01, workstation-02).

Gateways / firewalls / WAN aggregation

Type Count Model / notes Role
ML110 Gen9 1 Dell PowerEdge ML110 Gen9, 8–12 GbE OPNsense/pfSense WAN aggregator — between 6–10 Spectrum cable modems (WAN) and redundant UDM Pro gateways (LAN). Multi-WAN load balance/failover. See ML110_OPNSENSE_PFSENSE_WAN_AGGREGATOR.md. Prerequisite: Migrate all Proxmox workloads off ML110 before repurpose.
UniFi Dream Machine Pro (UDM Pro) 2 UniFi UDM Pro Primary: 192.168.11.1; second: 192.168.11.2 (HA/failover). WAN uplink from ML110 OPNsense/pfSense. Port forward 76.53.10.36 → NPMplus, etc.

Switching

Type Count Model Notes
UniFi XG 10G 16-port 2 UniFi Switch XG (10GbE, 16 ports) 10 Gbps backbone for Proxmox inter-node and Ceph storage; dual-attach for HA. See 13_NODE_NETWORK_AND_CABLING_CHECKLIST.md.

Use for 13-node cluster:

  • Ceph cluster network: Dedicated VLAN or subnet across the 10G switches; connect all 13 R630s (and R750s if in Ceph) for backend traffic.
  • Management (1G): UDM Pro(s) and existing 1G LAN; workstations and management IPs on same subnet or VLAN.
  • Redundancy: Two XG switches allow dual-attach per node (one link per switch or LACP).

References

Reference in New Issue View Git Blame Copy Permalink