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:
co-authored by
Claude Sonnet 5
parent
33d6eee9cf
commit
a3377083aa
@@ -6,7 +6,6 @@ import (
|
||||
"fmt"
|
||||
"log/slog"
|
||||
"os/exec"
|
||||
"regexp"
|
||||
"strconv"
|
||||
"strings"
|
||||
)
|
||||
@@ -39,48 +38,7 @@ func enrichPCIeWithNVIDIA(devs []schema.HardwarePCIeDevice) []schema.HardwarePCI
|
||||
slog.Info("nvidia: enrichment skipped", "err", err)
|
||||
return enrichPCIeWithNVIDIAData(devs, nil, false)
|
||||
}
|
||||
devs = enrichPCIeWithNVIDIAData(devs, gpuByBDF, true)
|
||||
return enrichPCIeWithNVIDIANVLinks(devs)
|
||||
}
|
||||
|
||||
// enrichPCIeWithNVIDIANVLinks attaches per-link NVLink status (nvidia-smi
|
||||
// nvlink -s) and error counters (nvidia-smi nvlink -e) to each GPU's
|
||||
// HardwarePCIeDevice entry, keyed by the "nvidia_gpu_index" telemetry set by
|
||||
// enrichPCIeWithNVIDIAData. Independent of NVSwitch/fabric-manager detection
|
||||
// so it also covers direct GPU-to-GPU bridge boards with no switch present.
|
||||
func enrichPCIeWithNVIDIANVLinks(devs []schema.HardwarePCIeDevice) []schema.HardwarePCIeDevice {
|
||||
statusByGPU, statusErr := nvlinkStatusFn()
|
||||
if statusErr != nil {
|
||||
slog.Info("nvidia: nvlink -s unavailable, skipping nvlink enrichment", "err", statusErr)
|
||||
return devs
|
||||
}
|
||||
errorsByGPU, errorsErr := nvlinkErrorsFn()
|
||||
if errorsErr != nil {
|
||||
slog.Info("nvidia: nvlink -e unavailable", "err", errorsErr)
|
||||
}
|
||||
|
||||
for i := range devs {
|
||||
if devs[i].Telemetry == nil {
|
||||
continue
|
||||
}
|
||||
idx, ok := devs[i].Telemetry["nvidia_gpu_index"].(int)
|
||||
if !ok {
|
||||
continue
|
||||
}
|
||||
ports, ok := statusByGPU[idx]
|
||||
if !ok {
|
||||
continue
|
||||
}
|
||||
for j := range ports {
|
||||
if counters, ok := errorsByGPU[idx][ports[j].Index]; ok {
|
||||
ports[j].ReplayErrors = &counters.Replay
|
||||
ports[j].RecoveryErrors = &counters.Recovery
|
||||
ports[j].CRCErrors = &counters.CRC
|
||||
}
|
||||
}
|
||||
devs[i].NVLinks = ports
|
||||
}
|
||||
return devs
|
||||
return enrichPCIeWithNVIDIAData(devs, gpuByBDF, true)
|
||||
}
|
||||
|
||||
func hasNVIDIADevices(devs []schema.HardwarePCIeDevice) bool {
|
||||
@@ -327,107 +285,3 @@ func injectNVIDIATelemetry(dev *schema.HardwarePCIeDevice, info nvidiaGPUInfo) {
|
||||
dev.MaxLinkWidth = info.PCIeLinkWidthMax
|
||||
}
|
||||
}
|
||||
|
||||
var (
|
||||
nvlinkGPUHeaderRe = regexp.MustCompile(`^GPU (\d+):`)
|
||||
nvlinkSpeedLineRe = regexp.MustCompile(`^Link (\d+):\s*([\d.]+)\s*GB/s`)
|
||||
nvlinkInactiveRe = regexp.MustCompile(`^Link (\d+):\s*<inactive>`)
|
||||
nvlinkErrorCounterRe = regexp.MustCompile(`^Link (\d+):\s*(Replay|Recovery|CRC) Errors:\s*(\d+)`)
|
||||
)
|
||||
|
||||
// nvlinkErrorCounters holds the per-link error counters reported by
|
||||
// "nvidia-smi nvlink -e" for one GPU.
|
||||
type nvlinkErrorCounters struct {
|
||||
Replay, Recovery, CRC int64
|
||||
}
|
||||
|
||||
// nvlinkStatusFn and nvlinkErrorsFn are swappable for testing.
|
||||
var (
|
||||
nvlinkStatusFn = queryNVIDIANVLinkStatusByGPU
|
||||
nvlinkErrorsFn = queryNVIDIANVLinkErrorsByGPU
|
||||
)
|
||||
|
||||
// queryNVIDIANVLinkStatusByGPU runs "nvidia-smi nvlink -s" and returns each
|
||||
// GPU's NVLink ports keyed by GPU index (as printed in the "GPU N:" header,
|
||||
// matching the index nvidia-smi --query-gpu also reports).
|
||||
func queryNVIDIANVLinkStatusByGPU() (map[int][]schema.HardwareNVLinkPort, error) {
|
||||
out, err := exec.Command("nvidia-smi", "nvlink", "-s").Output()
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return parseNVIDIANVLinkStatusByGPU(string(out)), nil
|
||||
}
|
||||
|
||||
func parseNVIDIANVLinkStatusByGPU(raw string) map[int][]schema.HardwareNVLinkPort {
|
||||
result := map[int][]schema.HardwareNVLinkPort{}
|
||||
currentGPU := -1
|
||||
for _, line := range strings.Split(raw, "\n") {
|
||||
trimmed := strings.TrimSpace(line)
|
||||
if m := nvlinkGPUHeaderRe.FindStringSubmatch(trimmed); m != nil {
|
||||
currentGPU, _ = strconv.Atoi(m[1])
|
||||
continue
|
||||
}
|
||||
if currentGPU < 0 {
|
||||
continue
|
||||
}
|
||||
if m := nvlinkInactiveRe.FindStringSubmatch(trimmed); m != nil {
|
||||
idx, _ := strconv.Atoi(m[1])
|
||||
result[currentGPU] = append(result[currentGPU], schema.HardwareNVLinkPort{Index: idx, Active: false})
|
||||
continue
|
||||
}
|
||||
if m := nvlinkSpeedLineRe.FindStringSubmatch(trimmed); m != nil {
|
||||
idx, _ := strconv.Atoi(m[1])
|
||||
port := schema.HardwareNVLinkPort{Index: idx, Active: true}
|
||||
if speed, err := strconv.ParseFloat(m[2], 64); err == nil {
|
||||
port.SpeedGBs = &speed
|
||||
}
|
||||
result[currentGPU] = append(result[currentGPU], port)
|
||||
}
|
||||
}
|
||||
return result
|
||||
}
|
||||
|
||||
// queryNVIDIANVLinkErrorsByGPU runs "nvidia-smi nvlink -e" and returns
|
||||
// per-link error counters keyed by GPU index then link index.
|
||||
func queryNVIDIANVLinkErrorsByGPU() (map[int]map[int]nvlinkErrorCounters, error) {
|
||||
out, err := exec.Command("nvidia-smi", "nvlink", "-e").Output()
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
return parseNVIDIANVLinkErrorsByGPU(string(out)), nil
|
||||
}
|
||||
|
||||
func parseNVIDIANVLinkErrorsByGPU(raw string) map[int]map[int]nvlinkErrorCounters {
|
||||
result := map[int]map[int]nvlinkErrorCounters{}
|
||||
currentGPU := -1
|
||||
for _, line := range strings.Split(raw, "\n") {
|
||||
trimmed := strings.TrimSpace(line)
|
||||
if m := nvlinkGPUHeaderRe.FindStringSubmatch(trimmed); m != nil {
|
||||
currentGPU, _ = strconv.Atoi(m[1])
|
||||
continue
|
||||
}
|
||||
if currentGPU < 0 {
|
||||
continue
|
||||
}
|
||||
m := nvlinkErrorCounterRe.FindStringSubmatch(trimmed)
|
||||
if m == nil {
|
||||
continue
|
||||
}
|
||||
linkIdx, _ := strconv.Atoi(m[1])
|
||||
count, _ := strconv.ParseInt(m[3], 10, 64)
|
||||
if result[currentGPU] == nil {
|
||||
result[currentGPU] = map[int]nvlinkErrorCounters{}
|
||||
}
|
||||
c := result[currentGPU][linkIdx]
|
||||
switch m[2] {
|
||||
case "Replay":
|
||||
c.Replay = count
|
||||
case "Recovery":
|
||||
c.Recovery = count
|
||||
case "CRC":
|
||||
c.CRC = count
|
||||
}
|
||||
result[currentGPU][linkIdx] = c
|
||||
}
|
||||
return result
|
||||
}
|
||||
|
||||
@@ -126,92 +126,3 @@ func TestEnrichPCIeWithNVIDIAData_driverMissingFallback(t *testing.T) {
|
||||
|
||||
func ptrInt64(v int64) *int64 { return &v }
|
||||
func ptrFloat(v float64) *float64 { return &v }
|
||||
|
||||
func TestParseNVIDIANVLinkStatusByGPU(t *testing.T) {
|
||||
// Real-world 2-GPU direct-bridge H100 SXM output: link 15 inactive on both GPUs.
|
||||
input := `GPU 0: NVIDIA H100 80GB HBM3 (UUID: GPU-a59f6931-c099-8fba-a0b3-08469d86f140)
|
||||
Link 0: 26.562 GB/s
|
||||
Link 15: <inactive>
|
||||
Link 17: 26.562 GB/s
|
||||
GPU 1: NVIDIA H100 80GB HBM3 (UUID: GPU-603fe750-0516-9db5-86ec-ea61af3fce35)
|
||||
Link 0: 26.562 GB/s
|
||||
Link 15: <inactive>
|
||||
`
|
||||
got := parseNVIDIANVLinkStatusByGPU(input)
|
||||
|
||||
if len(got[0]) != 3 {
|
||||
t.Fatalf("gpu0 ports=%d want 3 (%#v)", len(got[0]), got[0])
|
||||
}
|
||||
if got[0][1].Index != 15 || got[0][1].Active {
|
||||
t.Fatalf("gpu0 link15=%#v want inactive", got[0][1])
|
||||
}
|
||||
if got[0][0].SpeedGBs == nil || *got[0][0].SpeedGBs != 26.562 {
|
||||
t.Fatalf("gpu0 link0 speed=%#v want 26.562", got[0][0].SpeedGBs)
|
||||
}
|
||||
if len(got[1]) != 2 {
|
||||
t.Fatalf("gpu1 ports=%d want 2 (%#v)", len(got[1]), got[1])
|
||||
}
|
||||
if got[1][1].Active {
|
||||
t.Fatalf("gpu1 link15 should be inactive: %#v", got[1][1])
|
||||
}
|
||||
}
|
||||
|
||||
func TestParseNVIDIANVLinkErrorsByGPU(t *testing.T) {
|
||||
input := `GPU 0: NVIDIA H100 80GB HBM3 (UUID: GPU-a59f6931-c099-8fba-a0b3-08469d86f140)
|
||||
Link 0: Replay Errors: 0
|
||||
Link 0: Recovery Errors: 0
|
||||
Link 0: CRC Errors: 0
|
||||
Link 1: Replay Errors: 3
|
||||
Link 1: Recovery Errors: 1
|
||||
Link 1: CRC Errors: 2
|
||||
GPU 1: NVIDIA H100 80GB HBM3 (UUID: GPU-603fe750-0516-9db5-86ec-ea61af3fce35)
|
||||
Link 0: Replay Errors: 0
|
||||
Link 0: Recovery Errors: 0
|
||||
Link 0: CRC Errors: 0
|
||||
`
|
||||
got := parseNVIDIANVLinkErrorsByGPU(input)
|
||||
|
||||
c := got[0][1]
|
||||
if c.Replay != 3 || c.Recovery != 1 || c.CRC != 2 {
|
||||
t.Fatalf("gpu0 link1 counters=%#v want {3,1,2}", c)
|
||||
}
|
||||
zero := got[0][0]
|
||||
if zero.Replay != 0 || zero.Recovery != 0 || zero.CRC != 0 {
|
||||
t.Fatalf("gpu0 link0 counters=%#v want all zero", zero)
|
||||
}
|
||||
if _, ok := got[1][0]; !ok {
|
||||
t.Fatalf("expected gpu1 link0 entry present")
|
||||
}
|
||||
}
|
||||
|
||||
func TestEnrichPCIeWithNVIDIANVLinksAttachesPortsByIndex(t *testing.T) {
|
||||
oldStatus, oldErrors := nvlinkStatusFn, nvlinkErrorsFn
|
||||
t.Cleanup(func() { nvlinkStatusFn, nvlinkErrorsFn = oldStatus, oldErrors })
|
||||
|
||||
nvlinkStatusFn = func() (map[int][]schema.HardwareNVLinkPort, error) {
|
||||
return map[int][]schema.HardwareNVLinkPort{
|
||||
0: {{Index: 0, Active: true, SpeedGBs: ptrFloat(26.562)}, {Index: 15, Active: false}},
|
||||
}, nil
|
||||
}
|
||||
nvlinkErrorsFn = func() (map[int]map[int]nvlinkErrorCounters, error) {
|
||||
return map[int]map[int]nvlinkErrorCounters{
|
||||
0: {0: {Replay: 1}},
|
||||
}, nil
|
||||
}
|
||||
|
||||
devices := []schema.HardwarePCIeDevice{
|
||||
{Telemetry: map[string]any{"nvidia_gpu_index": 0}},
|
||||
}
|
||||
|
||||
out := enrichPCIeWithNVIDIANVLinks(devices)
|
||||
|
||||
if len(out[0].NVLinks) != 2 {
|
||||
t.Fatalf("nvlinks=%d want 2", len(out[0].NVLinks))
|
||||
}
|
||||
if out[0].NVLinks[0].ReplayErrors == nil || *out[0].NVLinks[0].ReplayErrors != 1 {
|
||||
t.Fatalf("link0 replay errors=%#v want 1", out[0].NVLinks[0].ReplayErrors)
|
||||
}
|
||||
if out[0].NVLinks[1].Active {
|
||||
t.Fatalf("link15 should stay inactive")
|
||||
}
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user