REPORT — v0.171.0: disk-health card device-model label (2026-07-25)

Overwritten per the standing rule. Pairs with agent v0.95.0 (SMART coverage); the R-50 island-bridge spike lives in felhom.eu.

Baseline → target

felhom-controller de14eed (v0.170.0) → committed f6a8249 (v0.171.0). Deployed to guest 9201.

What shipped (additive)

agentapi.SmartSummary gains ModelName (mirrors agent v0.95.0 model_name, omitempty).
diskDisplayLabel (disk_health.go) now prefers the device model over the raw storage name/UUID, falling back to Name (+ speed hint) on an older agent or a modelless disk. Old-agent payloads render exactly as before.
Files: internal/agentapi/client.go, internal/web/disk_health.go, +_test.go, CHANGELOG.

Tests + red-proof

go build/vet/test ./... — 27 packages pass, 0 fail; template/emoji/mojibake gates OK. TestDiskDisplayLabel_PrefersModel: model present → model label; old agent / nil model → Name(+hint) (A4 tolerance). Red-proof: drop the fallback (always return ModelName) → the old-agent/nil cases return "" and the test fails.

LIVE acceptance (guest 9201, with agent v0.95.0) — the operator's complaint reversed

Deployed 0.171.0 → Up (healthy). The "Lemezek állapota" card now shows BOTH physical disks with real verdicts + human model labels:

"AirDisk 512GB SSD" → Rendben, 34 °C (the system SSD — Fix A resolved the LVM root).
"TOSHIBA MQ04ABF100" → Rendben, 30 °C (the USB — Fix B + the model label).

No raw UUIDs, no "Nincs adat". /disks carries smart.health=PASSED + model_name for both (see the agent REPORT). The capability-map row was upgraded to PROVEN-LIVE (healthy path + delivery).

Observations

A live FAILING/Figyelmeztetés verdict still needs a genuinely degrading disk (both demo disks are healthy) — that arm stays unit + red-proof covered; the notification delivery path is separately live-proven (hub 400→200 + email, 2026-07-24).

2.0 KiB Raw Blame History