bee_layer_classify only ever tracked regular files and symlinks
(`find ... -type f -o -type l`), so any directory that is empty at
build time — like /var/log/nvidia-dcgm, correctly created and chowned
by the datacenter-gpu-manager postinst — was silently dropped from
every layer's rsync --files-from list and never reached the built
ISO. This is why bbc6fb1's 78d1b9b follow-up (seeding a marker file
just for that one path) kept the directory alive: it was a targeted
workaround for a general gap in the classifier, not a fix of it.
Replace that workaround with the general mechanism: classify also
walks every directory, computes the subset that is empty all the way
down (no file or symlink anywhere in its subtree — a directory that
does hold files needs no entry, rsync already recreates it as an
implied parent), and assigns each one to a layer via the same
dpkg-ownership / injected-rule precedence used for files. Add an
injected rule routing /var/log/nvidia-dcgm to 20-nvidia-platform,
alongside the DCGM binaries that actually use it, instead of letting
it fall through to base by default.
bee_layer_build folds each layer's empty-dir list into the same
rsync --files-from call; recursion into a directory that is
by-construction empty copies nothing extra. Revert the 9000/9999 hook
changes from 78d1b9b now that they're redundant, and cover the new
path with test-squashfs-layers.sh (ruled, unruled, and nested-empty
directories, asserted present in the merged rootfs after a real
mksquashfs/unsquashfs round-trip).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
bbc6fb1 chowned the directory correctly but it stayed empty at build
time, and bee_layer_classify (squashfs-layers.sh) only tracks regular
files and symlinks via `find ... -type f -o -type l` — an empty
directory is silently dropped from every layer's rsync --files-from
list and never reaches the built ISO. dcgmi diag's deployment check
still failed with DCGM_FR_FILE_CREATE_PERMISSIONS after rebuilding on
top of that fix because the directory simply didn't exist at boot.
Seed /var/log/nvidia-dcgm/.keep so the directory rides along the
classifier as an implied parent (rsync -a preserves its ownership),
and stop 9999-slim's log sweep from deleting that marker before the
classifier ever sees it.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
nv-hostengine drops privileges to --service-account nvidia-dcgm for all
file I/O and writes diagnostics to DCGM_HOME_DIR (/var/log/nvidia-dcgm,
per the packaged unit's Environment=). The package postinst doesn't
reliably create/own this directory inside a live-build chroot, so
dcgmi diag's software deployment check fails with
DCGM_FR_FILE_CREATE_PERMISSIONS at boot even though every GPU passes
the actual hardware tests.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
mirror.mephi.ru was timing out during lb binary_rootfs. Default to
deb.debian.org (Fastly-backed, geo-balanced) and allow overriding via
--mirror on build-in-container.sh/build.sh, or BEE_MIRROR in auto/config.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
On HGX B200 (Kaytus KR9288-X3) NVSwitch is managed via HMC/ERoT and never
appears on host lspci as 10de:0680, so the old check always returned
"not present" and silently skipped fabricmanager — leaving the fabric
untrained and CUDA/NCCL/dcgmi failing with "system not yet initialized"
on otherwise-healthy hardware. bee-check-nvswitch now checks, in order,
nvidia-smi's GPU Fabric GUID, /proc/driver/nvidia-nvswitch/devices, and
finally the legacy host-PCIe scan for older HGX generations.
- page_scenario.go: onclick built via JSON.stringify() embedded raw double
quotes inside a double-quoted HTML attribute, truncating the attribute so
the click handler never compiled; pass the name through an escaped
data-scenario-name attribute instead.
- build.sh: overlay staging rsyncs (OVERLAY_DIR->stage, stage->includes.chroot)
ran without --delete, so a scenario removed from the repo (a9924b0) stayed
baked into every ISO built from the persistent stage cache since — the
"second script" in the Scenario page's list.
- blackbox: rewritten around a deterministic local zip + incremental
patch-the-changed-suffix onto removable media, instead of walking/copying
~90 files through a synchronous ntfs-3g FUSE mount every cycle. journalctl
captures are now "--since last sync" (were "--since boot", growing with
uptime) and metrics.db is excluded (was copied whole every cycle).
- scenario: nvbandwidth-acs-ab now escalates GPU count (same-socket pair,
other socket's pair, one cross-socket pair, all GPUs) under each ACS state
instead of always running all 6 GPUs at once, using a new `bee
gpu-bandwidth-groups` subcommand that discovers socket layout from
`nvidia-smi topo -m` at runtime — gpu_indices is host-specific, so this
can't be baked into the scenario file.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
bee.nvidia.mode=gsp-off loads the driver with NVreg_EnableGpuFirmware=0 and
skips nvidia-uvm/nvidia-modeset, so CUDA (nvbandwidth/dcgm/nccl) can't run —
and on Hopper H200 GSP is effectively mandatory anyway, so the mode is dead
weight in this test image. Remove it from every boot entry. fail-safe is now
bee.nvidia.mode=normal pci=realloc iommu=pt, so the GPUs come up with uvm and
the GPU test actually runs under iommu=pt (the point of the entry). All four
sources (grub.cfg + live.cfg.in templates, build.sh write_canonical_*) match.
The internal GSP auto-fallback in bee-nvidia-load (normal mode, GSP init
hangs >90s → retry with EnableGpuFirmware=0, then still load uvm) is left
intact — it serves converted SXM->PCIe cards on other test targets and keeps
CUDA working, unlike the removed explicit menu mode.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
no-GUI is now just the normal entry minus the GUI: bee.gui=off with
bee.nvidia.mode=normal (drops the previous gsp-off/pci=realloc it had picked
up). fail-safe is now simply the GPU/PCIe-workaround boot —
bee.nvidia.mode=gsp-off pci=realloc iommu=pt over the standard base — instead
of the old noapic/nodma/nomce/nolapic/nosmp minimal mode, so it can actually
run GPU tests (nvbandwidth) to verify the iommu=pt half of the vendor fix.
Applied identically across all four sources (grub.cfg + live.cfg.in templates
and build.sh write_canonical_* generators).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The GRUB and isolinux menus (and their two sources of truth — the staged
templates and build.sh's write_canonical_* generators that rewrite the built
menu) had drifted: GRUB shipped a lean set, isolinux carried extra
display-debug entries (gsp-off, kms, kms-gsp-off) and lacked a WIPE entry in
the canonical writer. Collapse all four to one set:
normal, load-to-RAM (default), no-GUI/troubleshooting (gsp-off, pci=realloc),
fail-safe, WIPE ALL DISKS, Memory Test, + UEFI Firmware Settings (EFI only)
The fail-safe entry gains iommu=pt (AMD-Vi passthrough) — the OS-side half of
the vendor fix for the MSI CG480-S6053 cross-socket nvbandwidth reboot; ACS
can't be toggled from a stock-kernel cmdline so that stays a BIOS setting.
default stays index 1 = toram (config.cfg set default=1).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replaces the old nvbandwidth-all-gpu-power-watch scenario (dropped: the
power/IPMI-sampler framing was a dead-end — the blackbox shows every other
load test, incl. NCCL all-reduce across all 6 GPUs, passes with power fully
connected; only nvbandwidth on cross-socket GPU P2P reboots the box).
nvbandwidth-acs-ab is a two-phase reproduction for the MSI CG480-S6053 case.
Phase 1 disables PCIe ACS redirect at runtime via setpci (saving each
bridge's original ACS control word to /run/bee-acs-orig) then runs
nvbandwidth across all 6 GPUs — expected to survive. Phase 2 restores the
saved ACS values and repeats — expected to reproduce the reboot. Safe phase
first so its full blackbox reaches the USB stick before the risky phase can
reset the box (sync-bracket before the final job flushes phase-1 evidence).
ACS mutation runs via `bash -c` since scenario cmd is exec'd directly with no
shell. Both shipped copies (repo scenarios/ and image overlay) identical;
verified both parse via ParseScenarioJSON (4 jobs).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
`bee run <scenario>` called platform.System.RunScenario directly with an
empty base dir, which defaults to /var/log/bee-sat — a tree the blackbox
worker does not mirror (it mirrors DefaultExportDir, /appdata/bee/export).
So a scenario launched from the CLI wrote its logs where blackbox never
looked: the sync-bracket hooks fired and flushed the export dir, but the
scenario's own output never reached the USB stick. The web-UI path was
unaffected because it goes through App.RunScenario, which defaults the base
dir to DefaultSATBaseDir (under the export dir).
Route the CLI through App.RunScenario too — matching the doc comment that
already claimed both paths did — so `bee run` output lands under the mirrored
export tree and shows up in the blackbox capture.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The Scenario page's Run button enqueued a task with target "scenario" that
never launched any load: the queue runs each task in an external
bee-worker subprocess (RunPersistedTask -> executeTaskWithOptions in
task_runner.go), whose target switch had no "scenario" case, so the task
died with "unknown target: scenario". The case had only been added to
taskQueue.runTask's switch in tasks.go — a stale copy exercised solely by
unit tests, which is why the tests passed while the button did nothing.
Add the "scenario" case (ReadScenario -> ParseScenarioJSON -> RunScenario)
to executeTaskWithOptions, and collapse runTask into a thin delegate to it
so there is a single target dispatch. The old runTask switch had already
drifted into a subset (missing nvme-format, *-write, raid-*, nvidia-config,
...); removing it eliminates the divergence that hid this bug. Also drop the
now-orphaned taskQueue.statusDB helper.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- nvidia-smi underlines the topo -m header with ANSI CSI codes even when
writing to a file; both NVLink matrix parsers failed to find the header
and /topo showed "No NVLink-bonded GPU pairs found" on bonded systems.
- raid-lsi-create-mirror failed with "resources already in use" (exit 11)
on JBOD drives; the task now reads drive states and auto-converts
JBOD/UBad (set good force), releases hotspares, refuses Frgn/Onln with
actionable messages, and dumps preservedcache info when add vd fails.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Disk report now ends with a Conclusion section judging a drive NEW/USED
against loose thresholds (<110% capacity written, <210% read, <7d
uptime, <30 power cycles), listing which ones tripped. Data
Written/Read in the Usage section now scale to TB/PB via
formatBytesHuman instead of always printing raw GB. storageSATCommands
now runs smartctl with -i so SATA/SAS reports get Model/Serial/
Firmware/Capacity, which the Conclusion needs to evaluate the
write/read criteria (previously only -H -A was collected).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
nvme-cli emits large 64-bit counters as JSON-quoted strings on some
versions; the disk-report text generator only handled bare numbers and
{lo,hi} objects, so power_on_hours/data_units_read/data_units_written
etc. silently parsed as 0 while the structured collector path already
handled this correctly. Unify both paths on a single exported
JSONInt64/NVMeSmartLog/NVMeIDCtrl type in collector/storage.go instead
of keeping two independent nvme-cli JSON parsers in sync.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Show an explicit per-source status line after "Read All" instead of hiding
failed/blocked sources in a "(skipped: …)" tail. Sources blocked by a missing
Supermicro license (SFT-OOB-LIC / SFT-DCMS-SINGLE) are flagged in red with an
actionable message, so engineers see that SAA DMI is gated rather than silently
falling back to the futile ipmitool FRU path (BIOS re-syncs FRU from DMI on boot).
Also fix TestDashboardRendersRuntimeHealthTable, stale since 4f6579e moved
"inactive" to the OK service states: the fixture now uses a failed service and
the assertion matches the current contract (failed flagged, inactive not).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Tracks origin/main after rebase: adds per-column header filters for
severity in the viewer (feat(viewer): replace severity dropdown).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Picks up new contracts: hardware-ingest-json, submodule-integration,
go-database cursor safety, and several contract deduplication passes.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
A (hardware-ingest-json v2.8-2.9): remove sensor location fields from schema
and collector; tag HardwareMemory.Location as json:"-"; add PlatformConfig to
HardwareSnapshot.
B (no-hardcoded-vendors): consolidate PCI vendor IDs into collector/pci_vendors.go;
replace all vendor-name string checks in isGPUDevice, isNVIDIADevice, isMellanoxDevice,
isAMDGPUDevice, matchesGPUVendor (sat_overlay), and validateIsVendorGPU (page_validate)
with numeric vendor_id comparisons.
C (module-structure): split app/app.go (1413 lines) into app.go + app_format.go,
app_network.go, app_services.go, app_packs.go, app_install.go — no logic changes.
D (go-code-style): wrap bare return err in interfaceAdminState and
interfaceIPv4Addrs (platform/network.go) with fmt.Errorf context including
the interface name.
E (go-project-bible): add bible-local/architecture/data-model.md and
bible-local/architecture/api-surface.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Disabled PCIe devices (sysfs enable==0) carry no data traffic; their
link state has no operational impact. Switchtec PCIe switch management
endpoints on NVIDIA HGX H100 baseboards (and similar fabric controllers)
train at reduced speed intentionally and were producing spurious warnings.
Check is vendor-agnostic: reads enable attribute via existing helper,
no vendor/device ID hardcoding.
Documented in bible-local/decisions/2026-06-12-pcie-disabled-device-link-warning.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
nvme-cli emits smart-log counters as JSON strings and uses field names
avail_spare / percent_used instead of the prose names in the NVMe spec.
The nvmeSmartLog struct had int64 fields with wrong JSON tags — Unmarshal
returned an error and the whole health block was skipped, leaving every
NVMe drive with status=Unknown.
Fix: switch all numeric fields to jsonInt64 (already used for lsblk
block sizes) which accepts both bare numbers and quoted strings, and
correct the avail_spare / percent_used tag names.
Also fix validateIsVendorGPU for NVIDIA: previously counted any NVIDIA
PCIe device (including NVSwitch bridges) as a GPU, producing wrong
estimates (12 instead of 8 on an HGX H100 system). Now requires
device_class to be videocontroller or processingaccelerator, matching
the existing AMD filter logic.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
nvtop pulled nvidia-tesla-470-* via Recommends into the nogpu build.
Move it from bee.list.chroot into bee-nvidia and bee-amd lists so it
only appears in GPU variants.
Also remove the stray git-bible/ directory (was not gitignored) and
move grub-bitmap-error docs into bible-local/docs/.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Writing to /sys/class/scsi_host/hostX/scan on SAS controllers (e.g.
Adaptec smartpqi/PM8222-SHBA) triggers sas_user_scan which blocks
indefinitely, causing the audit to hang forever. Skip hosts that appear
under /sys/class/sas_host/ — SAS topology is discovered by the driver.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- storage: add jsonInt64 dual-format unmarshaler to handle lsblk output
change in util-linux 2.38 (LOG-SEC/PHY-SEC now emitted as JSON
integers, not quoted strings); fixes SATA disks invisible on Debian 12
- pcie: detect NVLink bridge mezzanine CX-7 cards (Mellanox x2, no host
net ifaces, DeviceName contains "NVLINK" in lspci -v) and mark them
with device_class="NVLinkBridge"; escalate PCIe link speed downgrade to
Critical for these cards (Gen3 on a fixed internal connector = hardware
fault, not a transient warning)
- pcie: cross-reference nvidia-smi topo to capture NVLink bond counts and
active status for all NVLink bridge cards
- packages: add infiniband-diags to ISO; provides ibstat required by
nvidia-fabricmanager-start.sh to enumerate IB devices before FM launch
(absence causes CUDA_ERROR_SYSTEM_NOT_READY)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Network: green if at least one interface has IPv4 (drop PARTIAL state).
Bee Services: treat inactive as OK — oneshot services (bee-sshsetup,
bee-preflight, bee-network, bee-audit, etc.) complete successfully and
exit to inactive; only failed is a real problem.
nvidia-fabricmanager: add ExecCondition=bee-check-nvswitch drop-in so
the service is silently skipped (inactive, not failed) on systems
without NVSwitch hardware (e.g. H200 NVL with direct NVLink, no
NVSwitch chips). bee-check-nvswitch detects NVSwitch via lspci
(vendor 10de, class 0680).
bee-nvidia.service: add ConditionPathExists=/usr/local/bin/bee-nvidia-load
so the unit is a no-op if somehow present in a non-nvidia build.
bee-boot-status: read /etc/bee-gpu-vendor and exclude bee-nvidia from
CRITICAL/ALL on non-nvidia builds, preventing boot hang if the unit
is unexpectedly present.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Convert bee-logo.png to RGBA (color type 6) and strip all metadata
chunks (cHRM, bKGD, tIME, tEXt) that confuse GRUB's minimal PNG parser
- Move terminal_output gfxterm before insmod png / theme load so the
theme initialises in an active gfxterm context
- Remove echo ASCII art banner from grub.cfg — with gfxterm active and
no terminal_box in the theme, echo output renders over the menu area
- Fix icon_heigh typo → icon_height; increase item_height 16→20 with
item_padding 0→2 for reliable text rendering in boot_menu
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- benchmark.go: retain sdrLastStep from final ramp step instead of
re-sampling after test when GPUs are already idle
- scripts/deploy.sh: build+deploy bee binary to remote host over SSH
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
DCMI reports only the managed power domain (~CPU+MB), missing GPU draw.
PSU AC input sensors cover full wall power. When samplePSUPower returns
data, sum the slots for PowerW; fall back to DCMI otherwise.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
\b does not fire between a digit and '_' because '_' is \w in RE2.
The pattern \bpsu?\s*([0-9]+)\b never matched PSU1_POWER_IN style
sensors, so parsePSUSDR (and PSUSlotsFromSDR / samplePSUPower) returned
empty results for MSI servers — causing all power graphs to fall back
to DCMI which reports ~half actual draw.
Added an explicit underscore-terminated pattern first in the list and
tests covering the MSI format.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
dcgmproftester and other GPU test subprocesses run inside the bee-web
cgroup and exceed 3G with 8 GPUs. OOM killer terminates the whole
service. No memory cap is appropriate on a LiveCD where GPU tests
legitimately use several GB.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Same fix as ramp steps: take sdrSingle snapshot after calibration
and prefer PSUInW over DCMI for singleIPMILoadedW. DCMI kept as
fallback. Log message indicates source.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When sdrStep.PSUInW is available, prefer it over DCMI for
ramp.ServerLoadedW and ServerDeltaW. DCMI on this platform (MSI 4-PSU)
reports ~half actual draw; SDR sums all PSU_POWER_IN sensors correctly.
Delta is now SDR-to-SDR (sdrStep.PSUInW - sdrIdle.PSUInW) for
consistency. DCMI path kept as fallback when SDR has no PSU data.
Log message now indicates the source (SDR PSU AC input vs DCMI).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
MSI servers place PSU_POWER_IN/OUT sensors on entity 3.0, not 10.N
(the IPMI "Power Supply" entity). The old parser filtered by entity ID
and found nothing, so the dashboard fell back to DCMI which reports
roughly half the actual draw.
Now delegates to collector.PSUSlotsFromSDR — the same name-based
matching already used in the Power Fit benchmark.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
A single dcgmproftester process without -i only loads GPU 0 regardless of
CUDA_VISIBLE_DEVICES. Now always routes multi-GPU runs through
bee-dcgmproftester-staggered (--stagger-seconds 0 for parallel mode),
which spawns one process per GPU so all GPUs are loaded simultaneously.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Increase stability profile duration from 33 min to 90 min by wiring
powerBenchDurationSec() into runBenchmarkPowerCalibration (was discarded)
- Collect per-step PSU slot readings, fan RPM/duty, and per-GPU telemetry
in ramp loop; add matching fields to NvidiaPowerBenchStep/NvidiaPowerBenchGPU
- Rewrite renderPowerBenchReport: replace Per-Slot Results with Single GPU
section, rework Ramp Sequence rows=runs/cols=GPUs, add PSU Performance
section (conditional on IPMI data), add transposed Single vs All-GPU
comparison table in per-GPU sections
- Add fmtMDTable helper (benchmark_table.go) and apply to all tables in
both power and performance reports so columns align in plain-text view
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Stability hardening (webui/app):
- readFileLimited(): защита от OOM при чтении audit JSON (100 MB),
component-status DB (10 MB) и лога задачи (50 MB)
- jobs.go: буферизованный лог задачи — один открытый fd на задачу
вместо open/write/close на каждую строку (устраняет тысячи syscall/сек
при GPU стресс-тестах)
- stability.go: экспоненциальный backoff в goRecoverLoop (2s→4s→…→60s),
сброс при успешном прогоне >30s, счётчик перезапусков в slog
- kill_workers.go: таймаут 5s на скан /proc, warn при срабатывании
- bee-web.service: MemoryMax=3G — OOM killer защищён
Build script:
- build.sh: удалён блок генерации grub-pc/grub.cfg + live.cfg.in —
мёртвый код с v8.25; grub-pc игнорируется live-build, а генерируемый
live.cfg.in перезаписывал правильный статический файл устаревшей
версией без tuning-параметров ядра и пунктов gsp-off/kms+gsp-off
- build.sh: dump_memtest_debug теперь логирует grub-efi/grub.cfg
вместо grub-pc/grub.cfg (было всегда "missing")
GRUB:
- live-theme/bee-logo.png: логотип пчелы 400×400px на чёрном фоне
- live-theme/theme.txt: + image компонент по центру в верхней трети
экрана; меню сдвинуто с 62% до 65%
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
As metrics.db grew (1 sample/5 s × hours), handleMetricsChartSVG called
LoadAll() on every chart request — loading all rows across 4 tables through a
single SQLite connection. With ~10 charts auto-refreshing in parallel, requests
queued behind each other, saturating the connection pool and pegging a CPU core.
Fix: add a background compactor that runs every hour via the metrics collector:
• Downsample: rows older than 2 h are thinned to 1 per minute (keep MIN(ts)
per ts/60 bucket) — retains chart shape while cutting row count by ~92 %.
• Prune: rows older than 48 h are deleted entirely.
• After prune: WAL checkpoint/truncate to release disk space.
LoadAll() in handleMetricsChartSVG is unchanged — it now stays fast because
the DB is kept small rather than capping the query window.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Snapshot IPMI "Power Supply" sensor states before and after each benchmark
run. Compare before/after to surface only *new* anomalies (pre-existing faults
are excluded). Results land in NvidiaBenchmarkResult.PSUIssues and
NvidiaPowerBenchResult.PSUIssues (JSON: psu_issues) and are printed in the
text benchmark report under a "PSU Issues" section.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- TestHandleAPIBenchmarkPowerFitRampQueuesBenchmarkPowerFitTasks: ramp-up
mode intentionally creates a single task (the runner handles 1→N internally
to avoid redundant repetition of earlier ramp steps). Updated the test to
expect 1 task and verify RampTotal=3 instead of asserting 3 separate tasks.
- TestBenchmarkPageRendersSavedResultsTable: benchmark page used "Performance
Results" as heading while the test looked for "Perf Results". Aligned the
page heading with the shorter label used everywhere else (task reports, etc.).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add BenchmarkEstimated* constants to benchmark_types.go from _v8 logs
(Standard Perf ~16 min, Standard Power Fit ~43 min, Stability Perf ~92 min)
- Update benchmark profile dropdown to show Perf / Power Fit timing per profile
- Add timing columns to Method Split table (Standard vs Stability per run type)
- Update burn preset labels to show "N min/GPU (sequential) or N min (parallel)"
- Clarify burn "one by one" description with sequential vs parallel scaling
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add SATEstimated* constants to sat.go derived from _v8 production logs,
with a rule to recalculate them whenever the script changes
- Extend validateInventory with NvidiaGPUCount to make estimates GPU-aware
- Update all validate card duration strings: CPU, memory, storage, NVIDIA GPU,
targeted stress/power, pulse test, NCCL, nvbandwidth
- Fix nvbandwidth description ("intended to stay short" → actual ~45 min)
- Top-level profile labels show computed total including GPU count
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add PSUReading struct and PSUs []PSUReading to LiveMetricSample
- Sample per-PSU input watts from IPMI SDR entity 10.x (Power Supply)
- Render stacked filled-area SVG chart (one layer per PSU, cumulative total)
- Fall back to single-line chart on systems with ≤1 PSU in SDR
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
grub-efi/grub.cfg: add KMS+GSP=off entry (was in isolinux, missing in GRUB)
isolinux/live.cfg.in: add full standard param set to all entries
(net.ifnames=0 biosdevname=0 mitigations=off transparent_hugepage=always
numa_balancing=disable nowatchdog nosoftlockup) to match grub-efi
bible-local/docs/iso-build-rules.md: add bootloader sync rule documenting
that grub-efi and isolinux must be kept in sync manually, listing canonical
entries and standard param set, and noting the grub-pc/grub-efi history.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Build uses --bootloaders "grub-efi,syslinux" so live-build reads
config/bootloaders/grub-efi/ for the UEFI GRUB config. The directory
was incorrectly named grub-pc, causing live-build to ignore our custom
grub.cfg and generate a default one (missing toram, GSP-off entries).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
live-boot already uses rsync --progress when /bin/rsync exists; without
it the copy falls back to silent cp -a. Add rsync to the ISO package
list and install an initramfs-tools hook (bee-rsync) that copies the
rsync binary + shared libs into the initrd via copy_exec. The hook then
rebuilds the initramfs so the change takes effect in the ISO's initrd.img.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- grub.cfg: add "load to RAM (toram)" entry to advanced submenu
- install_to_ram.go: resume from existing /dev/shm/bee-live copy if
source medium is unavailable after bee-web restart
- tasks.go: fix "Recovered after bee-web restart" shown on every run
(check j.lines before first append, not after)
- bee-install: retry unsquashfs up to 5x with wait-for-remount on
source loss; clear error message with bee-remount-medium hint
- bee-remount-medium: new script to find and remount live ISO source
after USB/CD reconnect; supports --wait polling mode
- 9000-bee-setup: chmod +x for bee-install and bee-remount-medium
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- NvidiaPowerBenchResult gains ServerPower *BenchmarkServerPower
- RunNvidiaPowerBench samples IPMI idle before Phase 1 and loaded via
background goroutine throughout Phase 2 ramp
- renderPowerBenchReport: new "Server vs GPU Power Comparison" table
with ratio annotation (✓ match / ⚠ minor / ✗ over-report)
- renderPowerBenchSummary: server_idle_w, server_loaded_w, server_delta_w,
server_reporting_ratio keys
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
400×400px PNG centered via feh --bg-center --image-bg '#000000'.
Fallback solid fill also changed to black.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace single aggregated badge per hardware category with individual
colored chips (O/W/F/?) for each ComponentStatusRecord. Added helper
functions: matchedRecords, firstNonEmpty. CSS classes: chip-ok/warn/fail/unknown.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
detectSlowdownTempExceedance scans steady-state metric rows per GPU and
emits a [WARNING] note + PARTIAL status if any sample >= SlowdownTempC.
Uses per-GPU threshold from nvidia-smi -q, fallback 80°C.
Distinct from p95-based TempHeadroomC check: catches even a single spike
above the slowdown threshold that would be smoothed out in aggregates.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Parse "GPU Shutdown Temp" and "GPU Slowdown Temp" from nvidia-smi -q verbose
output in enrichGPUInfoWithMaxClocks. Store as ShutdownTempC/SlowdownTempC
on benchmarkGPUInfo and BenchmarkGPUResult. Fallback: 90°C shutdown / 80°C
slowdown when not available.
TempHeadroomC = ShutdownTempC - P95TempC (per-GPU, not hardcoded 100°C).
Warning threshold: p95 >= SlowdownTempC. Critical: headroom < 10°C.
Report table shows both limits alongside headroom and p95 temp.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
All report strings are now English only.
Add detectPowerAnomaly: scans steady-state metric rows per GPU with a
5-sample rolling baseline; flags a sudden drop ≥30% while GPU usage >50%
as [HARD STOP] — indicates bad cable contact or VRM fault.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
CompositeScore = raw ComputeScore (TOPS). Throttling GPUs score lower
automatically — no quality multiplier distorting the compute signal.
Add ServerQualityScore (0-100): server infrastructure quality independent
of GPU model. Formula: 0.40×Stability + 0.30×PowerSustain + 0.30×Thermal.
Use to compare servers with the same GPU or flag bad server conditions.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
PowerSustainScore: power draw variance (CV) during load, not deviation from TDP.
ThermalSustainScore: temperature variance (CV) during load.
StabilityScore: fraction of time spent in thermal+power-cap throttling.
Remove NCCL bonus from quality_factor.
quality = 0.35 + 0.35×Stability + 0.15×PowerSustain + 0.15×ThermalSustain, cap 1.00.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Remove references to pre-benchmark power calibration and dcgmi
targeted_power. Document platform_power_score ramp-up methodology,
PowerSustainScore fallback to steady-state power, and full-budget
single-precision phases.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Power/Thermal Fit: cumulative fixed-limit ramp where each GPU's stable TDP
is found under real multi-GPU thermal load (all prior GPUs running at their
fixed limits). PlatformMaxTDPW = sum of stable limits across all GPUs.
Remove PlatformPowerScore from power test.
Performance Benchmark: remove pre-benchmark power calibration entirely.
After N single-card runs, execute k=2..N parallel ramp-up steps and compute
PlatformPowerScore = mean compute scalability vs best single-card TOPS.
PowerSustainScore falls back to Steady.AvgPowerW when calibration absent.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
mkdir -p LOG_DIR before writing the optional step log so that a race
with cleanup_build_log (EXIT trap archiving the log dir) does not cause
a "Directory nonexistent" error during lb binary_checksums / lb binary_iso.
Also downgrade apt-get update failure to a warning so a transient mirror
outage does not block kernel ABI auto-detection when the apt cache is warm.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Benchmark page now shows two result sections: Performance (scores) and
Power / Thermal Fit (slot table). After any benchmark task completes
the results section auto-refreshes via GET /api/benchmark/results
without a full page reload.
- Power results table shows each GPU slot with nominal TDP, achieved
stable power limit, and P95 observed power. Rows with derated cards
are highlighted amber so under-performing slots stand out at a glance.
Older runs are collapsed in a <details> summary.
- memtester is now wrapped with timeout(1) so a stuck memory controller
cannot cause Validate Memory to hang indefinitely. Wall-clock limit is
~2.5 min per 100 MB per pass plus a 2-minute buffer.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Phase 1 now calibrates each GPU individually (sequentially) so that
PowerRealizationPct reflects real degradation from neighbour thermals and
shared power rails. Previously the baseline came from an all-GPU-together
run, making realization always ≈100% at the final ramp step.
Ramp step 1 reuses single-card calibration results (no extra run); steps
2..N run targeted_power on the growing GPU subset with derating active.
Remove OccupiedSlots/OccupiedSlotsNote fields and occupiedSlots() helper —
they were compensation for the old all-GPU calibration approach.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Previously each GPU was calibrated sequentially (one card fully done
before the next started), producing the staircase temperature pattern
seen on the graph.
Now all GPUs run together in a single dcgmi diag -r targeted_power
session per attempt. This means:
- All cards are under realistic thermal load at the same time.
- A single DCGM session handles the run — no resource-busy contention
from concurrent dcgmi processes.
- Binary search state (lo/hi) is tracked independently per GPU; each
card converges to its own highest stable power limit.
- Throttle counter polling covers all active GPUs in the shared ticker.
- Resource-busy exponential back-off is shared (one DCGM session).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Remove telemetry-guided initial candidate; use strict binary search
midpoint at every step. Clean and predictable convergence in O(log N)
attempts within the allowed power range [minLimitW, startingLimitW].
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Power calibration previously stepped down 25 W at a time (linear),
requiring up to 6 attempts to find a stable limit within 150 W range.
New strategy:
- Binary search between minLimitW (lo, assumed stable floor) and the
starting/failed limit (hi, confirmed unstable), converging within a
10 W tolerance in ~4 attempts.
- For thermal throttle: the first-quarter telemetry rows estimate the
GPU's pre-throttle power draw. nextLimit = round5W(onset - 10 W) is
used as the initial candidate instead of the binary midpoint, landing
much closer to the true limit on the first step.
- On success: lo is updated and a higher level is tried (binary search
upward) until hi-lo ≤ tolerance, ensuring the highest stable limit is
found rather than the first stable one.
- Let targeted_power run to natural completion on throttle (no mid-run
SIGKILL) so nv-hostengine releases its diagnostic slot cleanly before
the next attempt.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
During power calibration: if a thermal throttle (sw_thermal/hw_thermal)
causes ≥20% clock drop while server fans are below 98% P95 duty cycle,
record a CoolingWarning on the GPU result and emit an actionable finding
telling the operator to rerun with fans manually fixed at 100%.
During steady-state benchmark: same signal enriches the existing
thermal_limited finding with fan duty cycle and clock drift values.
Covers both the main benchmark (buildBenchmarkFindings) and the power
bench (NvidiaPowerBenchResult.Findings).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
When a targeted_power attempt is cancelled (e.g. after sw_thermal
throttle), nv-hostengine holds the diagnostic slot asynchronously.
The next attempt immediately received DCGM_ST_IN_USE (exit 222)
and incorrectly derated the power limit.
Now: exit 222 is detected via isDCGMResourceBusy and triggers an
exponential back-off retry at the same power limit (1s, 2s, 4s, …
up to 256s). Once the back-off delay would exceed 300s the
calibration fails, indicating the slot is persistently held.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
mchus
