research(stp): a modelled Catalyst runs Rapid PVST+, so the word is rstp not ieee #2008
No reviewers
Labels
No milestone
No project
No assignees
2 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
jaylouisw/packetforge!2008
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "rev/1118-1701"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
One narrow question: is
ieeethe mode a modelled Catalyst is in before any configuration, and what does a real box print onshow spanning-tree's first line?Finding. Cisco documents the default mode per platform and release: the 2960-X is PVST+ at 15.2(2)E but Rapid PVST+ from 15.2(4)E, and the 9200 ships in Rapid PVST+. A box in Rapid PVST+ prints
Spanning tree enabled protocol rstp;ieeeis the PVST+ word. PacketForge printsieeefor all 19 Catalyst 9000 / 2960-X entries and times every port at 2x forward delay (30 s), so both halves of the answer describe a mode Cisco stopped defaulting to.The trap is that RSTP re-uses the same hello/max-age/forward-delay numbers (the 9200 default table prints 2 s / 20 s / 15 s), so neither a timer test nor the convergence figure can tell the modes apart - only the protocol word can, and nothing pins it (
test/l2.test.js:205only checks that the command runs).No engine change. The remedy is a second protocol mode - RSTP transition rules, per-VLAN instances and a
spanning-tree modecommand - which is a feature. Swapping the word alone would make the console claim rstp while converging like 802.1D. The note records what is known, what a boot capture would settle (timing on a 9200 edge port; whetherspanning-tree mode pvstflips the word), and why the catalogue cannot answer it per device: a device entry carries no software release.Also recorded, not acted on: the same 9200 table shows the long path-cost method as its default (1 Gb = 20000), while
engine/stp.jsimplements the short table - a separate question.Docs only:
research/stp-mode-default.md(117 lines, 11 sources).Gates:
check-control-bytesclean;syntax-check419 files OK; fullnpm test345 suites, 11982 assertions, 0 failed.4f1eaba7ea5fe86b1db95fe86b1db9to984a23f53cDispatch gap solved — close/reopen re-emits the event and CI runs. Merged as
f9b18c28.The missing checks were not event loss on the branch: run 6481 shows this PR's first SHA
4f1eaba7did get 8 checks, thenpr-rebase.ymlrebased onto the new main and force-pushed984a23f5, and the synchronize event for that workflow push never produced a run. A same-token no-op push reproduces the silence; a close + reopen (PATCH state=closed -> open) does not — it re-emitspull_request opened, and run 6491 built and passed 8/8 on984a23f5within 90 s.Workaround for the sibling loops: if a rebased PR sits with zero checks, close and reopen it rather than force-pushing again.