Lab Storage And Device Visibility Evidence
Date: 2026-06-20
Timezone: America/St_Johns
Supersession note: this is historical visibility evidence from before the
2026-06-21 internal production cutover. Current CT 121 state is public-safe
0.2.3; see [REDACTED:high-entropy].md.
Scope
Ran non-destructive checks to clarify remaining blockers after NUC-IPEX,
MacBook-Air, and Zenbook metadata surveys. This did not mutate production CT
121, sandbox CT 126, DNS, TLS, firewall, reverse proxy, public routes, VM
storage, snapshots, or device filesystems.
Proxmox Storage Readout
Host: 10.0.0.250
Hostname: Director
Observed pvesm status:
| Storage | Type | Status | Total KiB | Used KiB | Available KiB | Used |
|---|---|---|---|---|---|---|
| local | dir | active | 98,497,780 | 16,718,380 | 76,729,852 | 16.97% |
| local-lvm | lvmthin | active | 365,760,512 | 105,558,483 | 260,202,028 | 28.86% |
Observed LVM state:
- VG
pvesize: 475.94 GiB. - VG
pvefree: 16.00 GiB. - Thin pool
data: 348.82 GiB. - Thin pool data usage: 28.86%.
- Thin pool metadata usage: 1.29%.
- Snapshot LVs observed: 7.
- CT
126snapshots observed: migration-ready-20260620-081634mr022-20260620-190655mr022g-20260620-191542
Interpretation:
- The thin pool itself is not close to full by data or metadata percentage.
- The VG has limited unallocated free space, and multiple snapshot/thin volumes exist.
- Continue treating new snapshot-heavy work as a maintenance risk until an operator reviews storage policy, old snapshots, backups, and free-space targets.
Proxmox Container Readout
Observed pct list:
| CTID | Status | Name |
|---|---|---|
| 104 | running | shape-shop |
| 115 | stopped | aux |
| 116 | stopped | hedge |
| 117 | stopped | hedge-firm |
| 121 | running | uai-explorer |
| 123 | running | beacon-next-sandbox |
| 124 | running | cortex-sandbox-vnext |
| 125 | running | uai-bridge |
| 126 | running | uai-explorer-sandbox |
| 253 | running | pihole |
No QEMU VMs were listed by qm list.
Device Visibility Readout
Local ARP and TCP checks showed these notable reachable services:
| Address | Observed Name Or Role | Signal |
|---|---|---|
10.0.0.83 |
scotts-MacBook-Air.local |
SSH open; surveyed successfully as non-root scott. |
10.0.0.103 |
engineering |
SSH open as root; lab jump host. |
10.0.0.105 |
wow-weekend |
SSH open as root; not classified as UAI Explorer survey target. |
10.0.0.108 |
satisfactory |
SSH open as root; not classified as UAI Explorer survey target. |
10.0.0.121 |
uai-explorer |
App 8080 healthy on production 0.1.0; not mutated. |
10.0.0.126 |
uai-explorer-sandbox |
App 8080 ready on sandbox 0.2.2; raw survey outputs absent. |
10.0.0.248 |
forge |
SSH and Proxmox UI port open; root BatchMode hostname succeeded. |
10.0.0.249 |
echo |
SSH and Proxmox UI port open; root BatchMode hostname succeeded. |
10.0.0.250 |
Director |
SSH and Proxmox UI port open; current Proxmox host. |
10.0.0.253 |
likely infrastructure service | SSH, HTTP, and HTTPS open; BatchMode hostname was not established. |
Zenbook recheck:
scott@10.0.0.107SSH BatchMode timed out.- Ports
22,3389,5985, and445timed out. - Later 2026-06-20 recheck from this workstation showed ICMP loss, SSH timeout,
and a local neighbor state of
Unreachable. - The engineering jump host saw
10.0.0.107as a failed neighbor and returnedNo route to hostfor SSH. - Proxmox
Directoralso saw10.0.0.107as a failed neighbor; ICMP failed and SSH timed out. - After the operator re-suggested
10.0.0.107, a follow-up check still showed no local ARP entry, failed neighbor state from both the engineering jump host and ProxmoxDirector, 100% ICMP loss, and no TCP/22 route. This confirms the live-access issue was LAN visibility/current lease, not credentials. Zenbook is no longer a UAI Explorer blocker because the metadata survey and sanitized export are complete and no follow-up is required for this scope.
Dormant future Proxmox-node blocker:
- No authoritative hostname or IP for the dormant future Proxmox node was found in the current repo evidence.
- Several infrastructure-looking hosts are visible, but none should be assumed to be the dormant node without operator confirmation.
- The node still needs a metadata-only documentation survey before cluster join, disk reuse, or cleanup.
Follow-up identity survey:
[REDACTED:high-entropy].mdidentifies the online HomeLab Proxmox cluster nodes asDirector,echo, andforge.- It classifies
engineering,wow-weekend, andsatisfactoryas LXC guests onforge, and10.0.0.253as CTpiholeonDirector. - The dormant/future node remains unresolved because no separate powered-on host/IP was authoritatively identified.
Follow-up storage review:
[REDACTED:high-entropy].mdrecords a cluster-wide read-only storage review acrossDirector,echo, andforge.- The storage blocker is narrowed from a vague warning to an operator decision:
CT
126snapshot retention, archived CT snapshots, CT253pre-gpu-move,forgeCT103, andechoroot/template pressure need maintenance policy before snapshot-heavy work.