live-image: install stage tools and fail loudly if any are missing
CI / Lint + build + test (push) Successful in 1m32s
Release / release (push) Successful in 6m28s

The live image was still carrying the Phase 2 package list, so SMART,
CPUStress, and Network each hit a LookPath miss and returned
pass-with-skip. A run that skipped every real check still ended in
"completed" — nothing on the report said the image was broken.

Add smartmontools, stress-ng, fio, iperf3, lshw, lm-sensors,
e2fsprogs, and util-linux to mkosi.conf. Flip the three stages from
skip-pass to fail when their binary is missing so any future
packaging regression blocks the run instead of whispering past it.
Legitimate "no hardware" skips (no GPU, no hwmon, no disks,
non-destructive) are untouched.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
2026-04-18 16:39:28 -04:00
parent 0db790ae3e
commit e73e31af92
4 changed files with 40 additions and 10 deletions
+8 -4
View File
@@ -20,11 +20,15 @@ import (
// pages for the full duration, which is the Phase 4 health bar.
func CPUStress(ctx context.Context, d Deps) Outcome {
if _, err := exec.LookPath("stress-ng"); err != nil {
d.Warn("CPUStress: stress-ng not found in PATH — skipping stage")
// The live image ships stress-ng; absence is a packaging defect,
// not a benign local-dev scenario. Fail loudly so a regression
// in the image doesn't silently pass runs.
d.Error("CPUStress: stress-ng not found in PATH — live image is missing required tool")
return Outcome{
Passed: true,
Summary: "skipped (stress-ng missing)",
Extras: map[string]any{"skipped": true, "reason": "stress_ng_missing"},
Passed: false,
Message: "stress-ng binary missing from live image",
Summary: "failed (stress-ng missing)",
Extras: map[string]any{"reason": "stress_ng_missing"},
}
}
+6 -4
View File
@@ -24,11 +24,13 @@ type NetworkConfig struct {
// isn't reachable, or throughput is zero.
func Network(ctx context.Context, d Deps, cfg NetworkConfig) Outcome {
if _, err := exec.LookPath("iperf3"); err != nil {
d.Warn("Network: iperf3 not found — skipping stage")
// Live image ships iperf3; absence means packaging regression.
d.Error("Network: iperf3 not found — live image is missing required tool")
return Outcome{
Passed: true,
Summary: "skipped (iperf3 missing)",
Extras: map[string]any{"skipped": true, "reason": "iperf3_missing"},
Passed: false,
Message: "iperf3 binary missing from live image",
Summary: "failed (iperf3 missing)",
Extras: map[string]any{"reason": "iperf3_missing"},
}
}
host, err := deriveHost(cfg.OrchestratorURL)
+15
View File
@@ -21,6 +21,21 @@ import (
// surfaces as a per-disk "skipped" entry; the stage only fails if at
// least one disk reports !passed.
func SMART(ctx context.Context, d Deps) Outcome {
// smartctl absence is a packaging defect, not a per-disk skip. The
// per-disk `err != nil` path below catches "this device doesn't
// support SMART" (virtio-blk, exit 4); pre-checking the binary up
// front keeps that skip legitimate and fails the stage loudly if
// the live image lost its smartmontools package.
if _, err := exec.LookPath("smartctl"); err != nil {
d.Error("SMART: smartctl not found — live image is missing required tool")
return Outcome{
Passed: false,
Message: "smartctl binary missing from live image",
Summary: "failed (smartctl missing)",
Extras: map[string]any{"reason": "smartctl_missing"},
}
}
disks, err := listBlockDisks()
if err != nil {
d.Warn("SMART: failed to enumerate /sys/class/block: " + err.Error())
+11 -2
View File
@@ -46,6 +46,17 @@ Packages=
usbutils
initramfs-tools
zstd
# Stage binaries. Every package here backs a stage the agent runs —
# if any one goes missing the corresponding stage now fails the run
# (was: pass-with-skip). Keep this list in sync with agent/tests.
smartmontools
stress-ng
fio
iperf3
lshw
lm-sensors
e2fsprogs
util-linux
# Firmware. firmware-linux-nonfree on bookworm is a thin metapackage
# that does NOT pull i915 GuC/HuC — those live in firmware-misc-nonfree.
# Enumerate explicitly so the blob for whatever hardware we boot on
@@ -59,7 +70,5 @@ Packages=
amd64-microcode
firmware-linux-nonfree
# Phase 4 will add: smartmontools stress-ng fio iperf3 lshw lm-sensors
[Host]
# Copy the prebuilt Go agent in from the repo root via postinst.