🌱 Bump CAPI to v1.12.2, k8s to v1.34 and controller-runtime to v0.22.5#5857
Conversation
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
richardcase
left a comment
There was a problem hiding this comment.
@clebs - we'll also need to update the CAPI version in other places like the e2e config files. Like this: https://github.com/kubernetes-sigs/cluster-api-provider-aws/blob/main/test/e2e/data/e2e_conf.yaml#L30
|
/ok-to-test |
@richardcase I have updated CAPI and k8s versions based on previous bump PRs:
PTAL! |
b6a0781 to
b06184d
Compare
|
/retest |
1 similar comment
|
/retest |
|
/test pull-cluster-api-provider-aws-e2e-blocking |
|
/test pull-cluster-api-provider-aws-e2e-blocking |
|
@clebs this will need rebasing, then we can retest and see |
In CAPI v1.12 (v1beta2), the Cluster topology field changed from `spec.topology.class` (string) to `spec.topology.classRef.name`, and ClusterClass template references changed from `ref` (with apiVersion, kind, name) to `templateRef` (with kind, name only). Update all ClusterClass kustomize source templates to match the v1beta2 schema. Signed-off-by: Damiano Donati <damiano.donati@gmail.com>
In v1beta2, MachineDeploymentClass has bootstrap and infrastructure directly under the class entry, without the intermediate template wrapper that existed in v1beta1. Signed-off-by: Damiano Donati <damiano.donati@gmail.com>
The v1beta2 ClusterClassTemplateReference requires apiVersion in addition to kind and name. Signed-off-by: Damiano Donati <damiano.donati@gmail.com>
Kubernetes v1.34.8 includes a kubeadm change (kubernetes/kubernetes#138957) that removes the kubeadm:apiserver-kubelet-client ClusterRoleBinding on join. Without the matching CAPI fix (cluster-api#13664, backported in v1.12.8), the KCP controller gets Forbidden when proxying to etcd through the API server, blocking the pre-terminate kcp-cleanup hook indefinitely during rolling upgrades. Signed-off-by: Damiano Donati <damiano.donati@gmail.com>
The Flatcar stable AMI has a 15 GB root snapshot, which is larger than the 10 GB rootVolume size inherited from the default template. This causes EC2 instance creation to fail with "root volume size (10) must be greater than or equal to snapshot size (15)". Bump rootVolume to 16 GB in the ignition control-plane and worker AMI patches. Signed-off-by: Damiano Donati <damiano.donati@gmail.com>
…overlays The external-vpc-clusterclass and self-hosted-clusterclass kustomize overlays target cluster.x-k8s.io/v1beta1 for their JSON patches, but the Cluster template was bumped to v1beta2. This caused the patches to silently not apply, so variables like vpcID and selfHosted were missing from the generated templates. Signed-off-by: Damiano Donati <damiano.donati@gmail.com>
Signed-off-by: Damiano Donati <damiano.donati@gmail.com>
Kubernetes v1.34.2 introduced a kubeadm change (kubernetes/kubeadm#3294) that removed the implicit load balancer health gate from the wait-control-plane phase. This causes intermittent kubeadm join failures when the control plane endpoint is not yet fully converged. A fix has been merged upstream (kubernetes/kubernetes#139667) and will be included in v1.34.9. Until that release is available, temporarily pin workload clusters to v1.34.1 (pre-regression, with a published AMI) to unblock the CAPI v1.12 bump. The management cluster and EKS e2e configs are not affected.
… regression Kubernetes v1.34.x has a kubeadm bug (kubernetes/kubernetes#139667) where GetConfigMapWithShortRetry's 350ms polling budget is bypassed by a 10s http.Client.Timeout, causing intermittent worker join failures when the NLB has not fully converged. The fix is merged upstream but not yet released. Switch workload clusters to v1.33.4 (pre-regression) to unblock the CAPI v1.12 bump. Update etcd and coredns upgrade versions to match. Signed-off-by: Damiano Donati <damiano.donati@gmail.com>
The eks-nodeadm-clusterclass template and test were removed in 551148e but references remained in e2e_eks_conf.yaml and helpers.go, causing the EKS e2e suite to fail in BeforeSuite. Signed-off-by: Damiano Donati <damiano.donati@gmail.com>
Signed-off-by: Borja Clemente <borja.clemente@gmail.com>
|
/test pull-cluster-api-provider-aws-e2e-blocking pull-cluster-api-provider-aws-e2e pull-cluster-api-provider-aws-e2e-eks |
|
/test pull-cluster-api-provider-aws-e2e-eks Same unrelated flake from before, we can fix this in a follow up |
|
|
||
| // DNSResolutionCheck configures the behavior for checking the load balancer DNS resolution. | ||
| // Set to "None" to disable the check. | ||
| // If omitted, the provider will pick a reasonable default which may change over time. |
There was a problem hiding this comment.
Not sure can or should say that we will change the default behaviour.
|
/test pull-cluster-api-provider-aws-e2e-eks |
1 similar comment
|
/test pull-cluster-api-provider-aws-e2e-eks |
|
/lgtm 🎉 |
|
LGTM label has been added. DetailsGit tree hash: 3547b546020f4afb6f9c7d00844fbbd28b448cd3 |
|
/unhold |
What type of PR is this?
/kind support
What this PR does / why we need it:
This PR bumps CAPI to v1.12.2 and with it, k8s to v1.34.3 and controller-runtime to v0.22.5.
Needed to be up to date with CAPI releases.
Which issue(s) this PR fixes (optional, in
fixes #<issue number>(, fixes #<issue_number>, ...)format, will close the issue(s) when PR gets merged):Fixes #5834
Special notes for your reviewer:
None
TODO
This bump is affected by the following issues which needs to be addressed before we can successfully see this bump passing regular e2es:
Checklist:
Release note: