|
@@ -914,6 +914,15 @@ credential-free).
|
|
|
## Power
|
|
## Power
|
|
|
|
|
|
|
|
- `off`: SSH shutdown → wait → HA `turn_off` per-node entities + cluster extras
|
|
- `off`: SSH shutdown → wait → HA `turn_off` per-node entities + cluster extras
|
|
|
|
|
+- **shared extras are occupancy-guarded**: the fabric switch / fan are ONE
|
|
|
|
|
+ physical device listed in the `extra_entities` of several clusters. `off`
|
|
|
|
|
+ never turns off an extra that **another cluster with an open run-ledger entry
|
|
|
|
|
+ still references** (`runlock::open_runs`) — powering `8nodes-switch` down must
|
|
|
|
|
+ not kill a serve on `dgx58-switch` over the same `switch.sw6_mt`. A subset
|
|
|
|
|
+ `off --node` keeps ALL of the cluster's extras (its own remaining nodes need
|
|
|
|
|
+ the fabric). The ledger over-reports (a stale open run keeps the device on) —
|
|
|
|
|
+ the safe direction. Logs `keeping <entity> on — still in use by an active run
|
|
|
|
|
+ on cluster(s): …` (`src/power/toggle.rs::extras_held_by_other_clusters`)
|
|
|
- `on`: HA `turn_on` with **ensure-powered**: a node the caller believes down (web: last
|
|
- `on`: HA `turn_on` with **ensure-powered**: a node the caller believes down (web: last
|
|
|
fleet snapshot; CLI: 1.5 s TCP probe) whose plug reads on gets a **30 s plug power cycle**
|
|
fleet snapshot; CLI: 1.5 s TCP probe) whose plug reads on gets a **30 s plug power cycle**
|
|
|
(a shorter cut leaves the PSU rail energized — the machine stays in soft-off). Each node's
|
|
(a shorter cut leaves the PSU rail energized — the machine stays in soft-off). Each node's
|
