Primary Proxmox node for NPM, Vaultwarden, Homelab-MCP, Sentinel, and supporting infrastructure containers.
| Field | Value |
|---|---|
| Audience | Administrator |
| Public URL | https://proxmox2.pittsfamily.me |
| SSH | ssh proxmox2 |
| LAN IP | 10.0.2.13 |
| Version | pve-manager/9.1.9/ee7bad0a3d1546c9, kernel 7.0.2-2-pve |
| Hardware | Intel N150, 4 CPUs, 16 GB RAM |
| Last verified | 2026-06-12 |
proxmox2 is the currently verified Proxmox node for the AI/control-plane services. It hosts NPM, Pi-hole, Vaultwarden, Homelab-MCP, Authentik, and Sentinel. Public traffic reaches services through Cloudflare Tunnel to NPM, then to the internal service.
This is an administration node. Family users should not log in to Proxmox.
https://proxmox2.pittsfamily.me.SSH Key - Proxmox2 - root for SSH key reference and the relevant Proxmox account item if needed.ssh proxmox2
Do not SSH directly into LXC containers. Use pct exec from proxmox2.
Verified LXC inventory:
| CTID | Name | IP | Status | CPU | Memory | Disk | Role |
|---|---|---|---|---|---|---|---|
| 101 | npm | 10.0.2.14 | running | 2 cores | 512 MB | 8 GB | Nginx Proxy Manager, cloudflared |
| 102 | pihole | 10.0.2.15 | running | 1 core | 512 MB | 4 GB | DNS |
| 103 | mcp-server | 10.0.2.16 | stopped | 1 core | 512 MB | 4 GB | Decommissioned old MCP |
| 104 | vaultwarden | 10.0.2.17 | running | 1 core | 512 MB | 8 GB | Password vault |
| 105 | agent-memory-core | 10.0.2.18 | running | 2 cores | 2 GB | 16 GB | Homelab-MCP |
| 106 | ai-gateway | 10.0.2.24 | running | 2 cores | 4 GB | 32 GB | Authentik |
| 107 | sentinel | 10.0.2.25 | running | 2 cores | 4 GB | 32 GB | Sentinel OAuth AS and MCP proxy |
Storage verified:
| Storage | Type | Purpose | Approx. used |
|---|---|---|---|
local |
dir | ISO, templates, backup/import | 9.4 GB of 100.9 GB |
local-lvm |
lvmthin | LXC root disks and images | 37.5 GB of 374.5 GB |
Mnemosyne-Node2 |
NFS | Backups | 582.6 GB of 28.0 TB |
ssh proxmox2 'pveversion'
ssh proxmox2 'pvesh get /nodes/proxmox2/status --output-format json'
ssh proxmox2 'pct list'
ssh proxmox2 'pct status 107'
ssh proxmox2 'pct exec 107 -- bash -lc "hostname && docker ps"'
ssh proxmox2 'pct shutdown 107'
ssh proxmox2 'pct start 107'
Only stop containers during an approved maintenance window.
Use the Proxmox UI when possible. CLI pattern:
ssh proxmox2
apt update
apt list --upgradable
apt full-upgrade
reboot
Reboot only after confirming hosted services can tolerate downtime.
ssh proxmox2 'pct exec 105 -- bash -lc "cd /opt/homelab-mcp/deploy && docker compose logs --tail 80"'
ssh proxmox2 'pct exec 107 -- bash -lc "cd /opt/sentinel/deploy && docker compose logs --tail 80"'
Backups should land on Mnemosyne-Node2. Before major changes, create a snapshot or backup from the Proxmox UI, then verify the backup appears on the shared storage.
CLI checks:
ssh proxmox2 'pvesh get /nodes/proxmox2/storage --output-format json'
ssh proxmox2 'pct config 105'
Critical service state:
| CTID | Back up |
|---|---|
| 101 | /opt/npm/data, /opt/npm/letsencrypt, compose file |
| 104 | Vaultwarden data and config |
| 105 | Homelab-MCP Postgres volume, deploy config, secrets directory |
| 107 | Sentinel Postgres volume, sentinel-data volume, deploy config |
10.0.2.13 and gateway 10.0.2.1.Mnemosyne-Node2.| Issue | Current state | Action |
|---|---|---|
| CTID 103 is stopped | Old MCP container remains as stopped inventory | Keep stopped unless explicitly restoring old MCP for investigation. |
CTID 105 hostname is still agent-memory-core |
Service is now Homelab-MCP | Treat service identity as Homelab-MCP; rename later only with a maintenance plan. |
| Public services depend on NPM | NPM outage affects most URLs | Keep npm CTID 101 high priority in backups and recovery. |