Kindnet requires cluster to provide ipam via the node.spec.podcidr #17215
Conversation
k8s-ci-robot
added
cncf-cla: yes
This reverts commit ecadd7a. gce.UsesIPAliases(o) is assumed to be called after validating that cloud provider is GCP, with this commit the if statement applies to all cloud providers.
Change-Id: I655d0cdf245b5ab1610f7a856d9adb1326365090
|
/test pull-kops-e2e-gce-cni-kindnet |
k8s-ci-robot
added
the
lgtm
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: hakman The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
k8s-ci-robot
added
the
approved
|
/override pull-kops-e2e-gce-cni-kindnet |
|
@hakman: Overrode contexts on behalf of hakman: pull-kops-e2e-gce-cni-kindnet 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-sigs/prow repository. |
Different cloud providers provide the basic ipam functionality on the node.spec.podcidr via different components.
Typically this was done on the kube-controller-manager, but since the removal of the in-tree cloud provider code in Kubernetes, the GCP only ipam mode called CloudAllocator is not available, so it is provided by the cloud-controller-manager.
The commit in 27e8624 made a GCP only check to apply to all cloud providers, hence disabling the ipamm functionality in the KCM, and as a consequence the kindnet jobs that depend on that started failing
https://testgrid.k8s.io/kops-distro-u2404#kops-aws-cni-kindnet
Revert that PR and keep the dumping functionality that is unrelated to the IPAM functionality