The call for Contributor Summit Sessions is still open, for your discussions, workshops, or SIG meetings.
The legacy Linux package repositories (apt.kubernetes.io and yum.kubernetes.io AKA packages.cloud.google.com) have been frozen starting from September 13, 2023 and are going away in January 2024. Clusters running versions older than 1.24 and relying on the k8s project Linux package repositories (apt.kubernetes.io and yum.kubernetes.io AKA packages.cloud.google.com) will be affected. See the blog post.
WG-LTS upgrade survey is open until the end of January, please fill it out if you haven’t already.
Antonio Ojea is now a Tech Lead for SIG-Network.
Next Deadline: Production Readiness Review Freeze, February 1st
This PR aims to streamline the management of client-go features by registering them with the
DefaultFeatureGate instance. It establishes a connection that allows client-go to perform feature checks using the same instance used for Kubernetes feature checks. Notably, this integration ensures that client features seamlessly align with the existing
--feature-gates flag for Kubernetes components.
To circumvent a potential dependency cycle between component-base and client-go, this PR involves maintaining parallel definitions of feature types and constants in both. The successful integration hinges on the inclusion of an adapter to pass Kubernetes’ default feature gate instance to client-go.
spec.completions is an immutable field for Jobs with both
NonIndexed completion modes. This KEP proposes to allow mutating the
spec.completions field for
Indexed Jobs when
spec.completions equals to
spec.parallelism. With this, we can handle cases which require autoscaled indexed jobs. While such workloads can be modeled as StatefulSets, the Job API is better suited because of its features like allowing indexes to run to completion and better error handling. The KEP also describes in detail how the success and failure semantics are changed in the cases where
spec.completions is mutated.
This KEP is targeting to graduate to stable state in the upcoming v1.30 release.
kubectl drain won’t assume a pod is deleted after an API error
kubectl get cronjob shows timezone
HealthzBindAddress removed from SchedulerConfiguration API
Last Week In Kubernetes Development (LWKD) is a product of multiple contributors participating in Kubernetes SIG Contributor Experience. All original content is licensed Creative Commons Share-Alike, although linked content and images may be differently licensed. LWKD does collect some information on readers, see our privacy notice for details.
You may contribute to LWKD by submitting pull requests or issues on the LWKD github repo.