Infrastructure overview and topology

Proxmox infrastructure overview
Snapshot: 2026-09-13 UTC | Host: pve | Method: read-only host-local inspection via the connected OpenClaw node.
Summary
Proxmox VE 9.2.0 / pve-manager 9.2.11, kernel 7.0.14-15-pve. This is a single-node host with 24 logical CPUs, 62 GiB RAM, and an NVMe-backed LVM-thin datastore.
Live workload snapshot: 3 QEMU VMs (105, 6000, 6100) and 6 LXC containers (100, 101, 102, 107, 112, 119). Defined but not observed running: LXC 106 and 111. Storage volumes also indicate historical/stopped guests 104 and 110.
Guest inventory
QEMU VMs
ID
Name
Resources / notes
105
haos17-2
3 vCPU, 4 GiB; Home Assistant OS; USB passthrough
6000
OpenVino
4 vCPU, 17 GiB; PCI passthrough 00:02.0 and 00:0b.0; 150 GiB LVM disk
6100
claw
2 vCPU, 4.4 GiB; QEMU guest agent; qcow2 disk on /mnt/drive2
LXC containers
ID
Hostname
Limit / placement
State
100
npmplus
512 MiB; CPU 20
Running
101
vault.stargrid.com.au
1 GiB; CPU 20
Running
102
dock
4.63 GiB; CPUs 20-21; 4 CPU quota; nested Docker
Running
107
si.stargrid.com.au
2 GiB; CPUs 8,15
Running
112
sparky.stargrid.com.au
2 GiB; unpinned
Running
119
book.stargrid.com.au
1 GiB; unpinned
Running
106
zz.stargrid.com.au
2 GiB; CPUs 0,21
Defined; not seen running
111
stock.stargrid.com.au
3.95 GiB; CPUs 3,7
Defined; not seen running
Compute and storage
CPU: Intel Core Ultra 7 270K Plus, 24 logical CPUs, VT-x enabled.
Memory: 62 GiB total; ~34 GiB available at snapshot.
Swap: 8 GiB; ~3.2 GiB used.
System disk: Samsung 990 PRO 2 TB NVMe.
Root: ~94 GiB filesystem, ~57 GiB used (64%).
Proxmox LVM-thin pool: ~1.7 TiB.
Backup disk: 10.9 TB WD at /mnt/raidbackup1.
External RAID5: 3 × 10.9 TB WDC disks, ~21.8 TB md/LVM-backed Btrfs.
Temporary RAID0: 2 × 10.9 TB HGST disks, ~21.8 TB ext4 at /mnt/tempraid.
No ZFS pools detected.
Network and topology
LAN bridge: vmbr0 via nic0, host 192.168.1.115/24.
Tailscale: 100.87.74.45/32.
Wi-Fi present but down.
Guest veth/tap interfaces observed on vmbr0 for the live LXCs and VMs.
Proxmox firewall bridge devices observed for VM 6000, VM 6100 and CT 102.
LAN 192.168.1.0/24
        |
   nic0 / vmbr0 ---------------- Tailscale 100.87.74.45
        |
  +-----+-------------------------------+
  |                                     |
QEMU VMs                            LXC containers
105 haos17-2                         100 npmplus
6000 OpenVino                        101 vault.stargrid.com.au
6100 claw                            102 dock + Docker
                                     107 si.stargrid.com.au
                                     112 sparky.stargrid.com.au
                                     119 book.stargrid.com.au
        |
 NVMe LVM-thin datastore + external backup/RAID storage
Health findings and limitations
pve-cluster, pvestatd, pvedaemon and pveproxy were active.
corosync was inactive.
Local pvesh calls failed: 
Unable to load access control list: Unknown error -1.
Authenticated API checks returned HTTP 401 for version, nodes, guests, storage and health.
Therefore this is a host-local snapshot from running processes, LXC configs, block devices and interfaces—not a successful official API inventory.
Single-node layout is a single point of failure.
VM 6000 PCI passthrough and VM 105 USB passthrough create tighter hardware coupling.
Recommended follow-up
Investigate inactive corosync and pmxcfs ACL failures; do not restart services blindly.
Verify the API token and ACL once the control plane is healthy.
Check SMART, mdadm, LVM and Btrfs health.
Confirm backups and restore tests for critical VMs/LXCs.
Re-run official Proxmox inventory after repair.
Read-only snapshot. No guests were started, stopped, deleted, modified or snapshotted.