webui: add /topo server topology page
New read-only visualization page: CPU sockets as anchor nodes, PCIe
devices (GPU/NIC/RAID) linked to their NUMA-affine socket with edge
color derived strictly from link_speed vs max_link_speed (not from the
device's own Status, which can also be overwritten by SAT-test results
on the same field), PSU/BMC as standalone boxes with no connecting
line, and a separate NVLink Topology card (live nvidia-smi topo -m /
nvlink -s/-e queries, not persisted to any contract).
GPU-GPU edges are drawn strictly from the actual bonded-pair list
parsed out of "nvidia-smi topo -m" (parseGPUPairAdjacency), not from
adjacent box position in the layout — an earlier ASCII mockup drew a
"chain" through unrelated GPUs, which a dedicated regression test now
guards against. A bonded pair spanning two different NUMA nodes is
flagged Warning on the edge and on both GPU boxes, per project
decision that this is an anomaly worth surfacing, not a neutral fact.
Zero changes to the ingest contract: this reverts the HardwareNVLinkPort/
HardwarePCIeDevice.NVLinks field shipped in v11.55 (33d6eee) along with
its collector/nvidia.go enrichment — that field risked a 400 from
Reanimator Core's strict decoder without an RFC, and isn't needed since
the topo page queries nvidia-smi directly instead of reading it from
audit.json. The v11.55 systemd fix (bee-nvidia.service ordering,
nv-hostengine restart) and the nvlink-status/-errors/dcgmi dumps in the
support bundle are untouched.
Also reorganizes support bundle collection per project convention:
system/ is now LiveCD-operational logs only (Xorg, services, console,
network/FS of the host itself); all server-hardware dumps (lspci,
NVIDIA/NVLink/DCGM, fabric manager, PCIe AER, ethtool, mstflint) move
to techdump/, deduplicating two entries already produced by
platform/techdump.go. Adds nvidia-bug-report.sh (previously only
collected inside the NVIDIA SAT pack) and lscpu to the always-on dump.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
Mikhail Chusavitin
co-authored by
Claude Sonnet 5
Claude Sonnet 5
parent
33d6eee9cf
commit
a3377083aa
@@ -178,63 +178,48 @@ type HardwareStorage struct {
|
||||
|
||||
type HardwarePCIeDevice struct {
|
||||
HardwareComponentStatus
|
||||
Slot *string `json:"slot,omitempty"`
|
||||
VendorID *int `json:"vendor_id,omitempty"`
|
||||
DeviceID *int `json:"device_id,omitempty"`
|
||||
NUMANode *int `json:"numa_node,omitempty"`
|
||||
TemperatureC *float64 `json:"temperature_c,omitempty"`
|
||||
PowerW *float64 `json:"power_w,omitempty"`
|
||||
LifeRemainingPct *float64 `json:"life_remaining_pct,omitempty"`
|
||||
LifeUsedPct *float64 `json:"life_used_pct,omitempty"`
|
||||
ECCCorrectedTotal *int64 `json:"ecc_corrected_total,omitempty"`
|
||||
ECCUncorrectedTotal *int64 `json:"ecc_uncorrected_total,omitempty"`
|
||||
HWSlowdown *bool `json:"hw_slowdown,omitempty"`
|
||||
BatteryChargePct *float64 `json:"battery_charge_pct,omitempty"`
|
||||
BatteryHealthPct *float64 `json:"battery_health_pct,omitempty"`
|
||||
BatteryTemperatureC *float64 `json:"battery_temperature_c,omitempty"`
|
||||
BatteryVoltageV *float64 `json:"battery_voltage_v,omitempty"`
|
||||
BatteryReplaceRequired *bool `json:"battery_replace_required,omitempty"`
|
||||
SFPPresent *bool `json:"sfp_present,omitempty"`
|
||||
SFPIdentifier *string `json:"sfp_identifier,omitempty"`
|
||||
SFPConnector *string `json:"sfp_connector,omitempty"`
|
||||
SFPVendor *string `json:"sfp_vendor,omitempty"`
|
||||
SFPPartNumber *string `json:"sfp_part_number,omitempty"`
|
||||
SFPSerialNumber *string `json:"sfp_serial_number,omitempty"`
|
||||
SFPWavelengthNM *float64 `json:"sfp_wavelength_nm,omitempty"`
|
||||
SFPTemperatureC *float64 `json:"sfp_temperature_c,omitempty"`
|
||||
SFPTXPowerDBM *float64 `json:"sfp_tx_power_dbm,omitempty"`
|
||||
SFPRXPowerDBM *float64 `json:"sfp_rx_power_dbm,omitempty"`
|
||||
SFPVoltageV *float64 `json:"sfp_voltage_v,omitempty"`
|
||||
SFPBiasMA *float64 `json:"sfp_bias_ma,omitempty"`
|
||||
BDF *string `json:"-"`
|
||||
DeviceClass *string `json:"device_class,omitempty"`
|
||||
Manufacturer *string `json:"manufacturer,omitempty"`
|
||||
Model *string `json:"model,omitempty"`
|
||||
LinkWidth *int `json:"link_width,omitempty"`
|
||||
LinkSpeed *string `json:"link_speed,omitempty"`
|
||||
MaxLinkWidth *int `json:"max_link_width,omitempty"`
|
||||
MaxLinkSpeed *string `json:"max_link_speed,omitempty"`
|
||||
SerialNumber *string `json:"serial_number,omitempty"`
|
||||
Firmware *string `json:"firmware,omitempty"`
|
||||
MacAddresses []string `json:"mac_addresses,omitempty"`
|
||||
Present *bool `json:"present,omitempty"`
|
||||
IOMMUGroup *int `json:"iommu_group,omitempty"`
|
||||
NVLinks []HardwareNVLinkPort `json:"nvlinks,omitempty"`
|
||||
Telemetry map[string]any `json:"-"`
|
||||
}
|
||||
|
||||
// HardwareNVLinkPort describes a single NVLink lane on a GPU, as reported by
|
||||
// "nvidia-smi nvlink -s" (speed/active state) and "nvidia-smi nvlink -e"
|
||||
// (per-link error counters). Only populated on GPU-class HardwarePCIeDevice
|
||||
// entries. An inactive link is not necessarily a fault: some GPU SKUs/boards
|
||||
// reserve lanes as standby failover paths by design.
|
||||
type HardwareNVLinkPort struct {
|
||||
Index int `json:"index"`
|
||||
Active bool `json:"active"`
|
||||
SpeedGBs *float64 `json:"speed_gbs,omitempty"`
|
||||
ReplayErrors *int64 `json:"replay_errors,omitempty"`
|
||||
RecoveryErrors *int64 `json:"recovery_errors,omitempty"`
|
||||
CRCErrors *int64 `json:"crc_errors,omitempty"`
|
||||
Slot *string `json:"slot,omitempty"`
|
||||
VendorID *int `json:"vendor_id,omitempty"`
|
||||
DeviceID *int `json:"device_id,omitempty"`
|
||||
NUMANode *int `json:"numa_node,omitempty"`
|
||||
TemperatureC *float64 `json:"temperature_c,omitempty"`
|
||||
PowerW *float64 `json:"power_w,omitempty"`
|
||||
LifeRemainingPct *float64 `json:"life_remaining_pct,omitempty"`
|
||||
LifeUsedPct *float64 `json:"life_used_pct,omitempty"`
|
||||
ECCCorrectedTotal *int64 `json:"ecc_corrected_total,omitempty"`
|
||||
ECCUncorrectedTotal *int64 `json:"ecc_uncorrected_total,omitempty"`
|
||||
HWSlowdown *bool `json:"hw_slowdown,omitempty"`
|
||||
BatteryChargePct *float64 `json:"battery_charge_pct,omitempty"`
|
||||
BatteryHealthPct *float64 `json:"battery_health_pct,omitempty"`
|
||||
BatteryTemperatureC *float64 `json:"battery_temperature_c,omitempty"`
|
||||
BatteryVoltageV *float64 `json:"battery_voltage_v,omitempty"`
|
||||
BatteryReplaceRequired *bool `json:"battery_replace_required,omitempty"`
|
||||
SFPPresent *bool `json:"sfp_present,omitempty"`
|
||||
SFPIdentifier *string `json:"sfp_identifier,omitempty"`
|
||||
SFPConnector *string `json:"sfp_connector,omitempty"`
|
||||
SFPVendor *string `json:"sfp_vendor,omitempty"`
|
||||
SFPPartNumber *string `json:"sfp_part_number,omitempty"`
|
||||
SFPSerialNumber *string `json:"sfp_serial_number,omitempty"`
|
||||
SFPWavelengthNM *float64 `json:"sfp_wavelength_nm,omitempty"`
|
||||
SFPTemperatureC *float64 `json:"sfp_temperature_c,omitempty"`
|
||||
SFPTXPowerDBM *float64 `json:"sfp_tx_power_dbm,omitempty"`
|
||||
SFPRXPowerDBM *float64 `json:"sfp_rx_power_dbm,omitempty"`
|
||||
SFPVoltageV *float64 `json:"sfp_voltage_v,omitempty"`
|
||||
SFPBiasMA *float64 `json:"sfp_bias_ma,omitempty"`
|
||||
BDF *string `json:"-"`
|
||||
DeviceClass *string `json:"device_class,omitempty"`
|
||||
Manufacturer *string `json:"manufacturer,omitempty"`
|
||||
Model *string `json:"model,omitempty"`
|
||||
LinkWidth *int `json:"link_width,omitempty"`
|
||||
LinkSpeed *string `json:"link_speed,omitempty"`
|
||||
MaxLinkWidth *int `json:"max_link_width,omitempty"`
|
||||
MaxLinkSpeed *string `json:"max_link_speed,omitempty"`
|
||||
SerialNumber *string `json:"serial_number,omitempty"`
|
||||
Firmware *string `json:"firmware,omitempty"`
|
||||
MacAddresses []string `json:"mac_addresses,omitempty"`
|
||||
Present *bool `json:"present,omitempty"`
|
||||
IOMMUGroup *int `json:"iommu_group,omitempty"`
|
||||
Telemetry map[string]any `json:"-"`
|
||||
}
|
||||
|
||||
type HardwarePowerSupply struct {
|
||||
|
||||
Reference in New Issue
Block a user