Proxmox

The single Proxmox VE node (pve) that hosts nearly every SuperHiTech VM and container, its storage pools, and the guests on it.

Almost everything else in this section is a guest on one Proxmox host. Leif, nvrbackup, SHTops, obs-01, the web server, the PBX, the docs site — they’re all VMs or LXC containers on the pve node. This is the layer underneath the layer.

Leif manages it through the pve_* family: pve_node_status, pve_vm_list, pve_lxc_list, pve_storage_list, pve_template_list, and the lifecycle / exec tools (pve_lxc_start, pve_lxc_shutdown, pve_lxc_exec, pve_vm_exec, etc.).

The node

FieldValue
Node namepve (single-node, no cluster)
Statusonline
Memory~157 GiB total
Kernel6.8.x-pve
Tool routingpve_node_status for health; pve_* for everything else

Storage pools

Six storage backends, ZFS for the bulk guest disks and a Proxmox Backup Server datastore riding on Storj.

StorageTypeCapacityContentNotes
pool_2zfspool~10.8 TiB (mostly free)images, rootdirPrimary headroom pool
pool_1zfspool~7.1 TiB (~1.7 TiB used)images, rootdirMain guest disk pool
Backuplvm~3.0 TB (~350 GB used)rootdir, images
local-lvmlvmthin~758 GiB (~81% used)rootdir, imagesWatch this one — getting full
localdir~100 GBvztmpl, iso, backupTemplates + ISOs
storj-pbspbs~86 GB (~13 GB used)backupPBS datastore on Storj (shared)

Backups themselves are handled by the pbs-01 container (Proxmox Backup Server, CT 114) writing to the storj-pbs datastore — Storj object storage is the off-box destination. See Service Map → Backup.

Virtual machines (QEMU)

VMIDNameStatusRole
108unms.super-ht.comrunningUISP / UNMS — UniFi/Ubiquiti network management
112FreePBX17runningCurrent FreePBX (VoIP) — see Network Topology
107docs.super-ht.comrunningInternal docs host
101pbx.super-ht.comstoppedLegacy PBX, superseded by FreePBX17
100sheldon-ub01stoppedUbuntu VM (idle)

Containers (LXC)

The LXC guests are where most of the documented SuperHiTech services actually run. Each row links to its detail page where one exists.

VMIDNameStatusRole
110leifrunningLeif FastMCP host (10.10.0.25) — see Hosts
103nvrbackuprunningPricing app + backup host (10.10.0.14, SSH :2223) — see Hosts
109shtopsrunningOps / monitoring host (10.10.0.24) — see SHTops
111obs-01runningObservability stack (10.10.0.148) — see obs-01
105logsrunningLibreNMS host (10.10.0.250)
102webserverrunningVirtualmin / Postfix mail (webserver.super-ht.com) — see Service Map
114pbs-01runningProxmox Backup Server → storj-pbs
113ollama-01runningLocal LLM (Ollama) host
106ubuntu-24stoppedUbuntu 24.04 base/template
104CT104stoppedUnused container

Reaching containers

There’s no dedicated *_execute_command for most of these — only Leif, nvrbackup, and SHTops have their own shell tools. For everything else (obs-01, logs, ollama-01, webserver, pbs-01), go through the Proxmox layer with pve_lxc_exec against the container’s vmid on node pve, the same pattern documented on the obs-01 page.

TargetHow to reach it
leif (CT 110)local_execute_command
nvrbackup (CT 103)remote_execute_command
shtops (CT 109)shtops_execute_command
Any other containerpve_lxc_exec (node pve, the container’s vmid)
  • Hosts — host inventory and tool routing rules
  • SHTops — the ops host (CT 109)
  • obs-01 — observability stack (CT 111)
  • Service Map — services and the tools that manage them
  • Network Topology — VPNs, public endpoints, VoIP

Last verified against live systems: 2026-06-01 — see doc freshness.