fix(topology): attach HPE DIMMs and RAID member disks; import easy-bee event logs

HPE DL380 Gen11 BEE-SP v14.23 blackbox bundle (ADL-077):
- topology: accept "PROC n" DIMM locators, show the storage controller
  filtered from the export by ADL-062 when disks resolve to it, resolve
  storcli EID:Slt drives via VD -> OS drive -> controller BDF, hide BMC VGA,
  label numeric PSU slots "PSU n"
- easy_bee: parse SEL / dmesg -T event times, keep message_id and
  component_ref, pass platform_config through
- exporter: map ipmi-sel -> bmc and dmesg/kernel -> host in event_logs
- privacy: treat journald/syslog kernel lines as noise (linux.it false positive)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Mikhail ChusavitinandClaude Opus 5.5 committed 2026-10-01 12:49:08 +03:00
1 parent 9f5f7a98be
commit ff5b99fa5c
15 files changed
+349 -42

No files matched your search

+1
View File
@@ -42,6 +42,7 @@ hardware.network_adapters
hardware.network_cards
hardware.power_supplies
hardware.firmware
hardware.platform_config
```
`network_cards` is legacy/alternate source data.
+1 -1
View File
@@ -59,7 +59,7 @@ serial from its socket, model, board serial, or another component; reject source
| Vendor ID | Input family | Notes |
|-----------|--------------|-------|
| `dell` | TSR ZIP archives | Broad hardware, firmware, sensors, lifecycle events |
| `easy_bee` | `bee-support-*.tar.gz` | Imports the embedded hardware snapshot and static DIMM/storage/NVIDIA evidence used to build `topology.json`; without `bee-audit.json` (audit unfinished at capture) falls back to the bundled `nvidia-bug-report.txt` (ADL-076) |
| `easy_bee` | `bee-support-*.tar.gz`, blackbox `.zip` | Imports the embedded hardware snapshot (inventory, `event_logs` with source-native SEL/dmesg timestamps, `platform_config`) and static DIMM/storage/storcli/NVIDIA evidence used to build `topology.json` (ADL-077); without `bee-audit.json` (audit unfinished at capture) falls back to the bundled `nvidia-bug-report.txt` (ADL-076) |
| `h3c_g5` | H3C SDS G5 bundles | INI/XML/CSV-driven hardware and event parsing |
| `h3c_g6` | H3C SDS G6 bundles | Similar flow with G6-specific files |
| `hpe_ilo_ahs` | HPE iLO Active Health System (`.ahs`) | Proprietary `ABJR` container with gzip-compressed `zbb` members; parser combines SMBIOS-style inventory strings and embedded Redfish storage JSON |
+39
View File
@@ -2233,3 +2233,42 @@ message.
sensors and firmware are absent because they exist only in the audit snapshot.
- The root cause is on the collection side: bee-support should wait for `bee-audit.service`
(or mark the bundle as incomplete), and the R7715 `ipmitool sensor` stall needs a timeout.
## ADL-077 — Topology shows the storage controller behind RAID member disks; easy-bee event logs and platform_config are imported
**Date:** 2026-10-01
**Context:** BEE-SP v14.23 blackbox bundle of an HPE ProLiant DL380 Gen11 (CZ2D2Q07DQ): one CPU,
one DIMM, a MegaRAID (MR408i-o) with a RAID1 of two NVMe drives. The topology page showed the DIMM
and both disks under "Other / no socket affinity", no RAID controller, and the iLO Matrox VGA as a
"GPU". Causes: HPE DIMM locators read `PROC 1 DIMM 10`, not `CPU1_…`; the controller is dropped
from `pcie_devices` by ADL-062 (no serial/part number), so `topology.Build` never saw it; RAID
member drives carry the storcli `EID:Slt` slot (`252:1`), while `storage-controllers.txt` only
maps OS block devices by HCTL. Separately, all 402 snapshot `event_logs` were lost on export
(sources `ipmi-sel`/`dmesg` unmapped, timestamps unparsed) and `platform_config` was never read.
**Decision:**
- `topology.Build` additionally receives the controllers ADL-062 filters out
(`exporter.OnboardControllersForTopology`). Such a controller becomes a node only when a disk
resolves to it; the Reanimator export itself is unchanged.
- A storcli physical drive resolves to its controller through static evidence only:
PD → virtual drive (`storcli64-vds.json`) → `OS Drive Name` → `storage-controllers.txt` → BDF.
Drives outside any VD (`storcli64-drives.json`) inherit the BDF their controller index got from
a sibling VD.
- The DIMM socket locator accepts `PROC n` / `Processor n` next to `CPU n`.
- A video-class function with a BMC VGA vendor ID (Matrox `102b`, ASPEED `1a03`) is not a
topology node. A numeric PSU slot is labelled `PSU n`.
- `easy_bee` parses the source-native `event_time` strings bee passes through (ipmitool SEL
`MM/DD/YY hh:mm:ss UTC`, `dmesg -T`); a SEL `Pre-Init <ticks>` record stays undated (ADL-070).
`message_id` and `component_ref` are kept. The exporter maps `ipmi-sel` → `bmc` and
`dmesg`/`kernel` → `host`. The snapshot's `is_active` is not imported: it is bee's
latest-record-per-sensor guess, not an active-alarm snapshot (ADL-072).
- `hardware.platform_config` is added to `HardwareConfig` and passed through to the export
unchanged; `easy_bee` is the first source.
- Privacy scan treats journald/syslog `… kernel: …` lines as kernel noise like `[ 12.34] …`
lines (the `pps_core … <giometti@linux.it>` copyright was reported as customer `linux.it`).
**Consequences:**
- The bundle renders CPU 1 → Memory, NIC ×4, RAID → Storage ×2, with no "Other" block.
- The MegaRAID is still absent from the exported `pcie_devices` (ADL-062 stands); a controller
with no VD and no HCTL-mapped disk stays invisible in the topology too.
- Bee runtime/service/interface state events remain UI-only (no contract source).
+2 -1
View File
@@ -57,7 +57,8 @@ via `parser.PrivacyScanEnabled()`.
- **Clean-token boundary**: an FQDN glued to surrounding identifier chars
(`auth.backend.gssapi.store-creds`) or Title-cased (`OS.It`) is a code
fragment, not a host.
- Kernel ring-buffer lines (`[ 8.07]...`, `[ 8.07][ T1]...`) and Go
- Kernel ring-buffer lines (`[ 8.07]...`, `[ 8.07][ T1]...`), the same
messages in journald/syslog form (`Oct 01 09:32:18 host kernel: ...`) and Go
stack-trace lines (`/go-mod/`, `file.go:123`) are skipped for FQDN/email.
- `resolv` `domain`/`search` values must contain a dot.
- IPv4: skipped on `#`/`;` comment lines and version/spec-clause lines