Ciprian Hacman

@danny-does-stuff Could you try adding `mode: "0444"` to the file asset? https://kops.sigs.k8s.io/cluster_spec/#mode

@eddycharly If you still need this and want it added back, please do so and /cc me for review.

@JJotah This looks more like a Kubernetes bug than kOps. Did you open an issue in k/k also? If so, could you reference it? Is there anything interesting in kubelet...

I suggest to try https://github.com/kubernetes/kops/releases/tag/v1.27.0-beta.3 with a newer Karpenter and see if you can still reproduce the issue.

> Depending on the region, containerd can be slow to download ([example](https://storage.googleapis.com/kubernetes-jenkins/logs/e2e-kops-aws-karpenter/1705616628923764736/artifacts/i-04cb3757a26791415/journal.log)) 1m8s to download a 30MB binary is slower than expected, especially for these c6g.xlarge control plane instances. I...

Should be pretty easy to fix on the kOps side, but I would also recommend to raise an issue with Karpenter.

I guess CRDs need updating: ``` I0208 05:04:17.581270 5067 addon.go:192] Applying update from "s3://k8s-kops-prow/e2e-pr16327.pull-kops-e2e-aws-karpenter.test-cncf-aws.k8s.io/addons/karpenter.sh/k8s-1.19.yaml" I0208 05:04:17.581294 5067 s3fs.go:378] Reading file "s3://k8s-kops-prow/e2e-pr16327.pull-kops-e2e-aws-karpenter.test-cncf-aws.k8s.io/addons/karpenter.sh/k8s-1.19.yaml" W0208 05:04:17.942636 5067 health.go:69] expected status.conditions to be list,...

/test pull-kops-kubernetes-e2e-ubuntu-gce-build

/test pull-kops-kubernetes-e2e-ubuntu-gce-build

Related to https://github.com/projectcalico/calico/issues/8541. We will wait for v3.27.3.