Merge debug/prod into single ISO build, fix NVIDIA module loading

## ISO build consolidation
- Remove separate debug/prod split: overlay-debug/, build-debug.sh,
  mkimg.bee_debug.sh, genapkovl-bee_debug.sh all deleted
- Single overlay: iso/overlay/ (was overlay-debug content)
- Single build script: build.sh (SSH, TUI, NVIDIA, vendor tools, bee-release)
- Single mkimage profile: bee (with dropbear, dialog, strace, gcompat, etc.)

## NVIDIA fixes
- Modules now stored at /usr/local/lib/nvidia/ instead of
  /lib/modules/<kver>/extra/nvidia/ — modloop squashfs mounts over that
  path at boot making overlay content there inaccessible
- bee-nvidia init: load via insmod (absolute path), not modprobe
- bee-nvidia init: create libnvidia-ml.so.1/libcuda.so.1 symlinks in /usr/lib/
- build-nvidia-module.sh: always install linux-lts-dev (not conditional) —
  stale 6.6.x headers caused wrong-kernel modules that never loaded at runtime
- build-nvidia-module.sh: create soname symlinks in cache
- KERNEL_VERSION in VERSIONS updated 6.6 → 6.12
- gcompat added to ISO packages (nvidia-smi is a glibc binary on musl Alpine)

## Service ordering
- bee-audit: add `after bee-nvidia` so NVIDIA enrichment always succeeds

## New tooling
- iso/builder/smoketest.sh: SSH smoke test for post-boot ISO validation
- iso/builder/build-gpu-burn.sh: builds gpu_burn vendor binary (CUDA 12.8+)
- vendor/gpu_burn included automatically if placed in iso/vendor/

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
Mikhail Chusavitin
2026-03-06 20:14:18 +03:00
parent 0907ba07c3
commit 1768bb58dd
24 changed files with 1296 additions and 261 deletions

176
iso/builder/smoketest.sh Normal file
View File

@@ -0,0 +1,176 @@
#!/bin/sh
# smoketest.sh — run on a live ISO via SSH to verify all critical components.
#
# Usage:
# ssh root@<ip> 'sh -s' < smoketest.sh
# or: scp smoketest.sh root@<ip>:/var/tmp/ && ssh root@<ip> sh /var/tmp/smoketest.sh
#
# Exit code: 0 = all required checks passed, 1 = at least one required check failed.
PASS=0
FAIL=0
WARN=0
ok() { echo "[ OK ] $*"; PASS=$((PASS+1)); }
fail() { echo "[ FAIL ] $*"; FAIL=$((FAIL+1)); }
warn() { echo "[ WARN ] $*"; WARN=$((WARN+1)); }
info() { echo "[ INFO ] $*"; }
echo "========================================"
echo " bee live ISO smoketest"
echo " host: $(uname -n)"
echo " kernel: $(uname -r)"
echo " date: $(date -u)"
echo "========================================"
echo ""
# --- kernel version ---
KVER=$(uname -r)
info "kernel: $KVER"
# --- PATH ---
echo "-- PATH & binaries --"
for tool in dmidecode smartctl nvme ipmitool lspci audit; do
if p=$(PATH="/usr/local/bin:$PATH" command -v "$tool" 2>/dev/null); then
ok "$tool found: $p"
else
fail "$tool: NOT FOUND"
fi
done
for tool in nvidia-smi gpu_burn; do
if p=$(PATH="/usr/local/bin:$PATH" command -v "$tool" 2>/dev/null); then
ok "$tool found: $p"
else
warn "$tool: NOT FOUND (optional but expected)"
fi
done
echo ""
echo "-- NVIDIA modules --"
KO_DIR="/usr/local/lib/nvidia"
if [ -d "$KO_DIR" ]; then
ko_count=$(ls "$KO_DIR"/*.ko 2>/dev/null | wc -l)
ok "NVIDIA ko dir exists: $KO_DIR ($ko_count .ko files)"
else
fail "NVIDIA ko dir missing: $KO_DIR"
fi
for mod in nvidia nvidia_modeset nvidia_uvm; do
if /sbin/lsmod 2>/dev/null | grep -q "^$mod "; then
ok "module loaded: $mod"
else
fail "module NOT loaded: $mod"
fi
done
echo ""
echo "-- nvidia-smi --"
if PATH="/usr/local/bin:$PATH" command -v nvidia-smi >/dev/null 2>&1; then
if PATH="/usr/local/bin:$PATH" nvidia-smi -L 2>/dev/null | grep -q "GPU"; then
gpu_count=$(PATH="/usr/local/bin:$PATH" nvidia-smi -L 2>/dev/null | grep -c "GPU")
ok "nvidia-smi: $gpu_count GPU(s) found"
else
fail "nvidia-smi: runs but no GPUs detected"
fi
else
fail "nvidia-smi: not found in PATH"
fi
echo ""
echo "-- lib symlinks --"
for lib in libnvidia-ml libcuda; do
if [ -f "/usr/lib/${lib}.so.1" ] || [ -L "/usr/lib/${lib}.so.1" ]; then
ok "/usr/lib/${lib}.so.1 exists"
else
fail "/usr/lib/${lib}.so.1 MISSING (nvidia-smi will fail)"
fi
done
echo ""
echo "-- gcompat (glibc compat for nvidia-smi) --"
if [ -L /lib64/ld-linux-x86-64.so.2 ] || [ -f /lib64/ld-linux-x86-64.so.2 ]; then
ok "gcompat: /lib64/ld-linux-x86-64.so.2 present"
else
fail "gcompat: /lib64/ld-linux-x86-64.so.2 MISSING — nvidia-smi will fail to exec"
fi
echo ""
echo "-- openrc services --"
for svc in bee-nvidia bee-network; do
if rc-service "$svc" status >/dev/null 2>&1; then
ok "service running: $svc"
else
fail "service NOT running: $svc"
fi
done
for svc in bee-audit-debug dropbear bee-sshsetup; do
if [ -f "/etc/init.d/$svc" ]; then
if rc-service "$svc" status >/dev/null 2>&1; then
ok "service running: $svc"
else
warn "service not running: $svc (may be one-shot)"
fi
fi
done
echo ""
echo "-- audit binary --"
AUDIT=/usr/local/bin/audit
if [ -x "$AUDIT" ]; then
ok "audit binary: present"
ver=$("$AUDIT" --version 2>/dev/null || "$AUDIT" version 2>/dev/null || echo "unknown")
info "audit version: $ver"
else
fail "audit binary: NOT FOUND at $AUDIT"
fi
echo ""
echo "-- audit last run --"
if [ -f /var/log/bee-audit.log ]; then
last_line=$(tail -1 /var/log/bee-audit.log)
info "last log line: $last_line"
if grep -q "audit completed" /var/log/bee-audit.log 2>/dev/null; then
ok "audit: completed successfully"
elif grep -q "audit started" /var/log/bee-audit.log 2>/dev/null; then
warn "audit: started but may not have completed"
fi
# check for nvidia enrichment
if grep -q "nvidia: enrichment skipped" /var/log/bee-audit.log 2>/dev/null; then
reason=$(grep "nvidia: enrichment skipped" /var/log/bee-audit.log | tail -1)
fail "audit: nvidia enrichment skipped — $reason"
else
ok "audit: nvidia enrichment OK"
fi
else
warn "audit: no log found at /var/log/bee-audit.log"
fi
echo ""
echo "-- network --"
if ip route show default 2>/dev/null | grep -q "default"; then
gw=$(ip route show default | awk '{print $3}' | head -1)
ok "default route: $gw"
else
fail "no default route"
fi
if ping -c1 -W3 1.1.1.1 >/dev/null 2>&1; then
ok "internet: reachable (1.1.1.1)"
else
fail "internet: unreachable"
fi
echo ""
echo "-- /etc/profile.d/bee.sh PATH --"
if grep -q "/usr/local/bin" /etc/profile.d/bee.sh 2>/dev/null; then
ok "/etc/profile.d/bee.sh exports /usr/local/bin"
else
fail "/etc/profile.d/bee.sh does not add /usr/local/bin to PATH"
fi
echo ""
echo "========================================"
echo " Results: OK=$PASS FAIL=$FAIL WARN=$WARN"
echo "========================================"
[ "$FAIL" -eq 0 ] && exit 0 || exit 1