K8s v1.22 throws error on eks:index:VpcCni

[dswartz-altana]

What happened?

Using v1.22 for the kubernetes version of an eks.Cluster fails on creating a VPC CNI.

Steps to reproduce

Creating an eks cluster in Python:

        eks.Cluster(
            f"eks-v122",
            args=eks.ClusterArgs(
                cluster_security_group=self.cluster_sg,
                create_oidc_provider=True,
                enabled_cluster_log_types=[
                    "api",
                    "audit",
                    "authenticator",
                    "controllerManager",
                    "scheduler",
                ],
                endpoint_private_access=True,
                endpoint_public_access=True,
                name="eks-v122",
                node_root_volume_encrypted=True,
                private_subnet_ids=private_subnet_ids,
                provider_credential_opts=eks.KubeconfigOptionsArgs(role_arn=role_arn),
                public_subnet_ids=public_subnet_ids,
                storage_classes={
                    "gp2": eks.StorageClassArgs(
                        type="gp2",
                        allow_volume_expansion=True,
                        default=True,
                        encrypted=True,
                    )
                },
                version="1.22",
                vpc_id=vpc_id,
            ),
            opts=pulumi.ResourceOptions(
                parent=self,
                provider=provider,
            ),
        )

Expected Behavior

EKS cluster created without error.

Actual Behavior

EKS cluster resource fails with the following output:

 +  eks:index:VpcCni dswartz-platform-eks-vpc-cni creating error: Command failed: kubectl apply -f /var/folders/fn/0mznbwhn32x9wbjh92gqwlvm0000gp/T/tmp-25317N6mhwwaj1Dqk.tmp
 +  eks:index:VpcCni dswartz-platform-eks-vpc-cni **creating failed** error: Command failed: kubectl apply -f /var/folders/fn/0mznbwhn32x9wbjh92gqwlvm0000gp/T/tmp-25317N6mhwwaj1Dqk.tmp

Versions used

CLI
Version 3.28.0
Go Version go1.18
Go Compiler gc

Plugins
NAME VERSION
auth0 2.7.0
aws 4.38.1
databricks 0.2.0
eks 0.37.1
kubernetes 3.17.0
python unknown
random 4.4.2
tls 4.1.0

Host
OS darwin
Version 12.2.1
Arch arm64

This project is written in python (/opt/homebrew/bin/python3 v3.9.12)

(redacted)

NAME VERSION
alt 0.1.0
autopep8 1.6.0
black 22.1.0
click-example-complex 1.0
databricks-cli 0.16.4
flake8 4.0.1
netaddr 0.8.0
pip 22.0.4
pulumi-random 4.4.2
pulumi-tls 4.1.0
setuptools 60.10.0
tox 3.25.0
wheel 0.37.1

Additional context

No response

Contributing

Vote on this issue by adding a 👍 reaction.
To contribute a fix for this issue, leave a comment (and link to your pull request, if you've opened one already).

[jaxxstorm]

I believe this is because of:

aws/amazon-vpc-cni-k8s#1074

pulumi-eks/nodejs/eks/cni/aws-k8s-cni.yaml

Line 188 in 9c7c064

apiVersion: apiextensions.k8s.io/v1beta1

[2fifty6]

Is there a workaround? Maybe a transformation or vpc cni config option setting?

[viveklak]

This is fixed in v0.38.0. Thanks for your patience.