Skip to content

[e2e] Centralize Kubernetes version#713

Merged
imeoer merged 2 commits intocontainerd:mainfrom
DataDog:fricounet/fix-k8s-e2e-tests
Feb 11, 2026
Merged

[e2e] Centralize Kubernetes version#713
imeoer merged 2 commits intocontainerd:mainfrom
DataDog:fricounet/fix-k8s-e2e-tests

Conversation

@Fricounet
Copy link
Contributor

@Fricounet Fricounet commented Feb 10, 2026

Overview

Please briefly describe the changes your pull request makes.

Pin kubectl and kind node image to a single KUBE_VERSION variable in kind.sh instead of fetching the latest version dynamically from dl.k8s.io/release/stable.txt at build time, which sometimes break CI like here.
Also try to address some of the other flakiness of the e2e k8s jobs I've seen.

Related Issues

Please link to the relevant issue. For example: Fix #123 or Related #456.

Change Details

Please describe your changes in detail:

  • Hardcode the version with a new KUBE_VER in the Dockerfile

  • Restore the --config flag on kind create cluster that was lost in commit 1535ce6 when the workflow was refactored from inline steps to kind.sh, and pass --image from the centralized version.

  • Remove the duplicated image: field from kind.yaml since it is now controlled by the CLI flag.

  • Use containerdConfigPatches instead of replacing the entire containerd config

  • Wait for the API server healthz after kubelet restart in the CRI auth path.

Test Results

If you have any relevant screenshots or videos that can help illustrate your changes, please add them here.

To illustrate the the kind config file wasn't used before:
- kind logs from before this PR showing the node image version used is 1.30.0 and not 1.30.2

Change Type

Please select the type of change your pull request relates to:

  • Bug Fix
  • Feature Addition
  • Documentation Update
  • Code Refactoring
  • Performance Improvement
  • Other (please describe)

Self-Checklist

Before submitting a pull request, please ensure you have completed the following:

  • I have run a code style check and addressed any warnings/errors.
  • I have added appropriate comments to my code (if applicable).
  • I have updated the documentation (if applicable).
  • I have written appropriate unit tests.

Pin kubectl and kind node image to a single KUBE_VERSION
variable in kind.sh instead of fetching the latest version
dynamically from dl.k8s.io/release/stable.txt at build
time, which sometimes break CI.

Restore the --config flag on `kind create cluster` that
was lost in commit 1535ce6 when the workflow was
refactored from inline steps to kind.sh, and pass
--image from the centralized version.

Remove the duplicated image: field from kind.yaml since
it is now controlled by the CLI flag.
@Fricounet Fricounet force-pushed the fricounet/fix-k8s-e2e-tests branch from 6194321 to 7266319 Compare February 10, 2026 17:14
Use containerdConfigPatches instead of replacing the
entire containerd config. The custom config was missing
critical settings like SystemdCgroup, preventing pods
from starting.

Wait for the API server healthz after kubelet restart
in the CRI auth path.
@Fricounet Fricounet marked this pull request as ready for review February 10, 2026 18:02
Copy link
Collaborator

@imeoer imeoer left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks, LGTM!

@imeoer imeoer merged commit d985907 into containerd:main Feb 11, 2026
17 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants