show power inline is missing: PoE is modelled and displayed, but the console cannot report it #1295

Closed
opened 2026-11-09 23:21:36 +00:00 by jaylouisw · 0 comments
Owner

PoE is modelled, displayed and tested — and the console cannot report it.

Simulator.computePoe() has allocated the budget since PoE was modelled: ports served in name order, a denial
emitting a poe-deny event with a reason, the powered device flagged unpowered. The inspector panel displays
the whole thing. The console — the surface this project is for — has no show that mentions it, so "how much
power is drawn, and why did that port get nothing" can only be answered by opening a browser.

Every other modelled mechanism has a show: interfaces, VLANs, MAC table, ARP, IP routes, VRRP, LLDP,
etherchannel, spanning tree, DHCP pools, DNS zones. PoE is the one that does not.

This is a read-only view of an existing result, not a new mechanism: no configuration surface, no new
allocation rule, and every figure taken from sim.poe.

What it must get right

sw1#show power inline
Interface  Admin  Oper  Draw     Powered device
1/1/1      up     on    12.95 W  PD1
1/1/2      up     off   12.95 W  PD2  (denied)
1/1/3      up     off   -        the far end is not a powered device
1/1/4      down   off   -        administratively down
22 ports with nothing plugged in: 1/1/5, 1/1/6, 1/1/7, 1/1/8, 1/1/9 (+17 more)

Total: 12.95 W of 25 W in use, 12.05 W left, 1 port denied
  1/1/2: PSE budget exhausted: 12.95 W requested, 12.05 W left
  • every PSE port is accounted for, including the ones the engine leaves out of its table — which are the ones
    an operator is usually asking about. The engine has no entry for a shut-down, unplugged, or non-PD port, so the
    console names it from the hardware and gives the reason in the same words show interfaces uses;
  • the denial sentence is the timeline's, not a re-rendering: the engine now keeps reason (and remaining)
    on the denied port entry as well as in the event, so the two cannot drift. The panel already counted denials
    with no reason at all;
  • idle ports are summarised, because a wall of nothing plugged in is how the one row that matters gets
    missed; any of them can still be asked for by name;
  • three things it refuses to fake:
    • no simulation run → Not measured (no simulation run): nothing has computed the power budget yet, not zeros;
    • PSE ports but no budget → the ports are listed and the budget is reported as not modelled, not 0 W it never
      measured;
    • a question about one port → no device-wide total, because the number would be about something else;
  • an unknown port name is refused by name, and a port that exists but cannot supply power says so.
PoE is modelled, displayed and tested — and the console cannot report it. `Simulator.computePoe()` has allocated the budget since PoE was modelled: ports served in name order, a denial emitting a `poe-deny` event with a reason, the powered device flagged `unpowered`. The inspector panel displays the whole thing. The console — the surface this project is for — has no `show` that mentions it, so "how much power is drawn, and why did that port get nothing" can only be answered by opening a browser. Every other modelled mechanism has a `show`: interfaces, VLANs, MAC table, ARP, IP routes, VRRP, LLDP, etherchannel, spanning tree, DHCP pools, DNS zones. PoE is the one that does not. This is a **read-only view of an existing result**, not a new mechanism: no configuration surface, no new allocation rule, and every figure taken from `sim.poe`. ## What it must get right ``` sw1#show power inline Interface Admin Oper Draw Powered device 1/1/1 up on 12.95 W PD1 1/1/2 up off 12.95 W PD2 (denied) 1/1/3 up off - the far end is not a powered device 1/1/4 down off - administratively down 22 ports with nothing plugged in: 1/1/5, 1/1/6, 1/1/7, 1/1/8, 1/1/9 (+17 more) Total: 12.95 W of 25 W in use, 12.05 W left, 1 port denied 1/1/2: PSE budget exhausted: 12.95 W requested, 12.05 W left ``` * **every PSE port is accounted for**, including the ones the engine leaves out of its table — which are the ones an operator is usually asking about. The engine has no entry for a shut-down, unplugged, or non-PD port, so the console names it from the hardware and gives the reason in the same words `show interfaces` uses; * **the denial sentence is the timeline's**, not a re-rendering: the engine now keeps `reason` (and `remaining`) on the denied port entry as well as in the event, so the two cannot drift. The panel already counted denials with no reason at all; * **idle ports are summarised**, because a wall of `nothing plugged in` is how the one row that matters gets missed; any of them can still be asked for by name; * **three things it refuses to fake**: - no simulation run → `Not measured (no simulation run): nothing has computed the power budget yet`, not zeros; - PSE ports but no budget → the ports are listed and the budget is reported as not modelled, not `0 W` it never measured; - a question about one port → no device-wide total, because the number would be about something else; * an unknown port name is refused by name, and a port that exists but cannot supply power says so.
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
jaylouisw/packetforge#1295
No description provided.