New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
kube-controller-manager 1.20.5 ttl_controller panic when calling patchNodeWithAnnotation #101045
Comments
@s1113950: This issue is currently awaiting triage. If a SIG or subproject determines this is a relevant issue, they will accept it by applying the The Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
/sig api-machinery |
The same error is fixed in #100013 on 1.21. |
Can confirm that after I applied the patch from #100013 the panics went away |
ttl_controller belongs to SIG Apps |
/assign |
Fix was backported to both 1.20 and 1.19. They should be available in the next minor releases 1.20.7 and 1.19.11. Those releases are planned for the next week 2021-05-12 https://github.com/kubernetes/sig-release/blob/master/releases/patch-releases.md#detailed-release-history-for-active-branches |
/close |
@serathius: Closing this issue. In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. |
What happened:
We (Palantir) are in the process of upgrading our k8s stack from 1.19.5 to 1.20.5, and are noticing panics in the
ttl_controller
every so often of the kube-controller-manager. We run 3 of them, and alerting is fine and controller-managers are able to start, but eventually they'll fail at a rate of ~100x in a 24 hr period.Sample logs after ~20 mins of running on a test stack, after the kube-controller-manager pod
successfully acquired lease kube-system/kube-controller-manager
, andtll_controller
tried to patch a node with an annotation:What you expected to happen:
kube-controller-manager with no panics
How to reproduce it (as minimally and precisely as possible):
We run hyperkube-based containers as static pods on AWS with the following flags on the controller-manager:
Anything else we need to know?:
We have our own fork of (mostly) vanilla k8s, and didn't experience this prior to upgrading to k8s 1.20.5. Th only difference between this version and 1.19 is we added the
"--logging-format=json"
flag.Environment:
kubectl version
):cat /etc/os-release
):uname -a
):Linux ip-10-0-2-51.ec2.internal 4.15.0-1097-aws #104-Ubuntu SMP Fri Mar 19 18:19:00 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux
The text was updated successfully, but these errors were encountered: