kubernetes: [Flaking Test] Conformance-GCE-master-kubetest2 (client rate limiter Wait returned an error)
Failure cluster c538bc0857d4164bb14a
Error text:
[FAILED] client rate limiter Wait returned an error: rate: Wait(n=1) would exceed context deadline
In [It] at: test/e2e/apimachinery/resource_quota.go:1191 @ 11/09/23 14:53:30.003
TestGrid: https://testgrid.k8s.io/sig-release-master-blocking#Conformance - GCE - master - kubetest2
Recent failures:
11/15/2023, 5:50:05 PM ci-kubernetes-e2e-gci-gce 11/15/2023, 1:48:47 PM ci-kubernetes-gce-conformance-latest-kubetest2 11/15/2023, 10:47:10 AM ci-kubernetes-gce-conformance-latest-kubetest2 11/15/2023, 4:20:04 AM ci-kubernetes-e2e-ubuntu-gce-network-policies 11/15/2023, 3:54:04 AM ci-cos-containerd-e2e-cos-gce
/kind flake
/sig api-machinery
About this issue
- Original URL
- State: closed
- Created 8 months ago
- Comments: 25 (25 by maintainers)
ok, my bad, I didn’t follow the code right, the resync period for the resource quota controller is an option in
https://github.com/kubernetes/kubernetes/blob/94f15bbbcbe952762b7f5e6e3f77d86ecec7d7c2/cmd/kube-controller-manager/app/core.go#L552
and the default is 5 minutes
https://github.com/kubernetes/kubernetes/blob/94f15bbbcbe952762b7f5e6e3f77d86ecec7d7c2/pkg/controller/resourcequota/config/v1alpha1/defaults.go#L40-L42
that matches with the observation on the failed test , everything adds up now
@aojea: Using assertions inside a
wait.Pollcallback as in https://github.com/kubernetes/kubernetes/blob/8254569d50d6bbe8d60bcdfb16eac337f6e60514/test/e2e/apimachinery/resource_quota.go#L1189-L1191 is wrong. An error should be returned instead. Or better yet, don’t usewait.Poll…An assertion inside a
gomega.Eventuallyis not ideal either, but I thinkgomega.Eventuallywould handle it better thanwait.Poll.