Skip to content

docs(snmp): the iLO's SNMPv1 is off, and four documents stop pointing at the issue for it - #585

Merged
Gerrrt merged 1 commit into
mainfrom
gerrrt/snmpv3-ilo-ups-authpriv-76f804
Sep 20, 2026
Merged

Gerrrt merged 1 commit into
mainfrom
gerrrt/snmpv3-ilo-ups-authpriv-76f804

Conversation

@Gerrrt

@Gerrrt Gerrrt commented Sep 20, 2026

Copy link
Copy Markdown
Owner

Follow-up to #581, which moved shiva to SNMPv3 authPriv but deliberately claimed nothing about SNMPv1: that half was unproved at the time, so the documents pointed at #85 for it. It is proved now.

What happened on the device

SNMPv1 Request off on the iLO, then ./scripts/snmp-verify.sh --old → rejected for shiva. That probe is always v2c whatever the device speaks now, so it is precisely the check that v1/v2c access is gone. A second --device shiva afterwards, because disabling v1 is another save of the same page:

Current community
  PASS shiva      10.0.30.10   "Integrated Lights-Out 4 2.82 Feb 06 2023"
Unauthenticated PDUs (junk community must be refused)
  PASS shiva      10.0.30.10   refuses a junk community over GET and GETBULK, short and long
Stock communities (must be refused)
  PASS shiva      10.0.30.10   refuses public private
all SNMP targets verified

What changed here

Documentation only — four passages that asserted less than what is true:

  • SECURITY.md known-exposure row: the BMC answers no community at all now.
  • docs/security.md SNMPv2c section: the lab-segment poll is not readable from the lab segment any more.
  • docs/hardware.md Management row.
  • docs/roadmap.md Move the iLO and the UPS card to SNMPv3 authPriv (ADR-0036) #85 bullet, which also gains the scrape measurement ADR-0036 asked for rather than assumed:
scrape_duration_seconds on shiva
v2c, 24 h average before 12.05 s
v3, 1 h average after 12.09 s

60 samples in that hour, no gaps, min_over_time(up{job="snmp"}[30m]) = 1 on all four devices.

Verified

make validate clean. The first local run failed observability: promtool check config with a spurious "file_sd target file does not exist" against a file that is present and tracked — the known local flake under a busy docker daemon; the rerun passed it, and this change is four markdown files.

#85 stays open: mjolnir has not moved yet.

🤖 Generated with Claude Code

… at the issue for it

§4.5 ran on `shiva` today: *SNMPv1 Request* disabled, and
`snmp-verify.sh --old` returned `rejected` — the iLO no longer answers
its own former community over v2c. Until now the documents claimed the
v3 poll and deferred the v1 half to #85, which was right while it was
unproved and is understating the estate now that it is not.

`SECURITY.md`'s known-exposure row, `docs/security.md`'s SNMPv2c
section, `docs/hardware.md`'s Management row and the roadmap's #85
bullet all say it instead.

The roadmap bullet also carries the scrape cost ADR-0036 asked to be
measured rather than assumed: 12.09 s averaged over the hour after the
move against 12.05 s over the day before it, 60 samples and no gaps.
The v3 discovery round trip is not visible against a 12-second walk.

`mjolnir` is still open, so #85 is not.

Refs #85

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant