Repository navigation
fix: nok8s preflight misses replica ports and Envoy admin, and passes silently without ss - #1755
Merged
Merged
Conversation
UgaTheDev
requested review from
Vezio,
achandrasekar,
kalantar,
maugustosilva,
mengmeiye and
namasl
as code owners
August 10, 2026 05:29
Collaborator
|
Hi @UgaTheDev The PR looks good to me! Can you resolve the conflicts so we'll approve and merge it? Thanks. |
Signed-off-by: Kush Zingade <[email protected]>
UgaTheDev
force-pushed
the
fix/nok8s-preflight-port-check
branch
from
August 11, 2026 06:10
66deac4 to
4e2a6ef
Compare
mengmeiye
approved these changes
Aug 11, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The nok8s preflight port check has three gaps, all of which let a standup proceed into a conflict it could have predicted.
hostPortis checked for replica 0 only. Withnok8s.vllm.replicas: Nthe workers bindhostPort .. hostPort+N-1(34_nok8s-containers.yaml.j2:24), so8001..800Nwere never looked at. Envoy's admin port 19000 is hard-coded in33_nok8s-envoy.yaml.j2and listed as a required free port indocs/nok8s.md:37, but was not in the list either.The third one is the reason I noticed. The check runs
ss -ltninsideif res.success and res.stdout:, so on any host without iproute2 it reports nothing and passes. On macOSssdoes not exist, so the check has never done anything there.tests/test_nok8s_plan.py:157-170only covered the path wheresssucceeds and finds no listeners, so nothing caught it.This expands the port list, adds
lsof -nP -iTCP -sTCP:LISTENas a fallback probe so the check works on macOS, and warns when no probe is available rather than passing silently. A warning that says it could not check is worth more than a silent pass that reads as all clear.Four regression tests, each verified to fail before the change: every replica port checked, admin port checked, lsof fallback used when ss is absent, and the no-probe-available warning. Full suite matches the pre-existing baseline on main (same 4 unrelated failures in
test_config_schema.py).Fixes #1703