kubernetes: Flake: [sig-network] Network should set TCP CLOSE_WAIT timeout

Which jobs are failing: pull-kubernetes-e2e-kops-aws

Which test(s) are failing: [sig-network] Network should set TCP CLOSE_WAIT timeout Test (that’s what the name appears as, I don’t have any more info)

Since when has it been failing: 11-28-2018

Testgrid link: https://gubernator.k8s.io/build/kubernetes-jenkins/pr-logs/pull/71479/pull-kubernetes-e2e-kops-aws/115150/ https://testgrid.k8s.io/presubmits-kubernetes-blocking#pull-kubernetes-e2e-kops-aws&include-filter-by-regex=^Overall%24|Test

Reason for failure:

[sig-network] Network should set TCP CLOSE_WAIT timeout

Expected error:
    <*strconv.NumError | 0xc001812c90>: {
        Func: "Atoi",
        Num: "",
        Err: {s: "invalid syntax"},
    }
    strconv.Atoi: parsing "": invalid syntax
not to have occurred
test/e2e/network/kube_proxy.go:204

Test

error during ./hack/ginkgo-e2e.sh --ginkgo.flakeAttempts=2 --ginkgo.skip=\[Slow\]|\[Serial\]|\[Disruptive\]|\[Flaky\]|\[Feature:.+\]|\[HPA\]|Dashboard|Services.*functioning.*NodePort --report-dir=/workspace/_artifacts --disable-log-dump=true: exit status 1

Anything else we need to know: Filing this as a flake as subsequent test executions succeeded.

/kind failing-test /kind flake

About this issue

  • Original URL
  • State: closed
  • Created 6 years ago
  • Comments: 23 (16 by maintainers)

Most upvoted comments

I’m going to close this because there are no errors on the pull-kubernetes-e2e-gce-* jobs in the last 5 days, and only one error in the ci-kubernetes-e2e-gci-gce-*that I can’t explain, because theoretically, the loop to find the close_wait connection has to run every second during 30 seconds and, in that failure, it only runs 3 times.

There are other errors of this test with the output pod ran to completion, but those jobs have a considerable amount of test failed (>100), so I’m concerned if that can be something environmental causing these failures.

/close

feel free to reopen

They all pull from ci/latest

Just to use the latest failure listed on triage as an example: https://prow.k8s.io/view/gcs/kubernetes-jenkins/logs/ci-kubernetes-e2e-gci-gce-coredns-nodecache/1252809431276589059

v1.19.0-alpha.2.8+85ee5fdd901619

Looking at the commits in #90278 (eg: https://github.com/kubernetes/kubernetes/pull/90278/commits/638e077b4b93753a97315dcfb9c88f424d33f333) I see that it landed before v1.19.0-alpha.2, so your fix is in there