[Bug] Error: no output "VPC" in stack of the load balancer iamserviceaccount when creating nodegroup

[NicolasGrosjeanProbayes]

What were you trying to accomplish?

I wanted to change the AMI family of the nodegroup, so I have deleted the nodegroup with eksctl delete nodegroup then creating again with eksctl create nodegroup.

What happened?

It works well on my dev cluster but not in the production cluster, I have the following error

Error: no output "VPC" in stack "eksctl-prod-1-cluster-addon-iamserviceaccount-kube-system-aws-load-balancer-controller-prod-1-cluster"

Sometimes the error is

Error: no output "SecurityGroup" in stack "eksctl-prod-1-cluster-addon-iamserviceaccount-kube-system-aws-load-balancer-controller-prod-1-cluster"

How to reproduce it?

Like I am unable to reproduce it on the dev cluster, I don't know how to reproduce it.

In Cloudformation, we can see the following steps concerning the prod cluster.

Logs

$ eksctl create nodegroup   --cluster ${AWS_CLUSTER_NAME}   --region ${AWS_REGION}   --name test   --node-ami-family Bottlerocket   --node-type g4dn.xlarge   --node-volume-size 35   --nodes 0   --nodes-min 0   --nodes-max 10 -v 4
2024-11-26 15:32:06 [▶]  Setting credentials expiry window to 30 minutes
2024-11-26 15:32:06 [▶]  role ARN for the current session is REDACTED
2024-11-26 15:32:07 [▶]  cluster = &types.Cluster{AccessConfig:(*types.AccessConfigResponse)(REDACTED), Arn:(*string)(REDACTED), CertificateAuthority:(*types.Certificate)(REDACTED), ClientRequestToken:(*string)(nil), ConnectorConfig:(*types.ConnectorConfigResponse)(nil), CreatedAt:time.Date(2024, time.September, 23, 11, 44, 39, 13000000, time.UTC), EncryptionConfig:[]types.EncryptionConfig(nil), Endpoint:(*string)(REDACTED), Health:(*types.ClusterHealth)(REDACTED), Id:(*string)(nil), Identity:(*types.Identity)(REDACTED), KubernetesNetworkConfig:(*types.KubernetesNetworkConfigResponse)(REDACTED), Logging:(*types.Logging)(REDACTED), Name:(*string)(REDACTED), OutpostConfig:(*types.OutpostConfigResponse)(nil), PlatformVersion:(*string)(REDACTED), ResourcesVpcConfig:(*types.VpcConfigResponse)(REDACTED), RoleArn:(*string)(REDACTED), Status:"ACTIVE", Tags:map[string]string{"Name":"eksctl-prod-1-cluster-cluster/ControlPlane", "alpha.eksctl.io/cluster-name":"prod-1-cluster", "alpha.eksctl.io/cluster-oidc-enabled":"true", "alpha.eksctl.io/eksctl-version":"0.185.0", "aws:cloudformation:logical-id":"ControlPlane", "aws:cloudformation:stack-id":"arn:aws:cloudformation:eu-west-1:REDACTED:stack/eksctl-prod-1-cluster-cluster/REDACTED", "aws:cloudformation:stack-name":"eksctl-prod-1-cluster-cluster", "eksctl.cluster.k8s.io/v1alpha1/cluster-name":"prod-1-cluster"}, Version:(*string)(REDACTED), noSmithyDocumentSerde:document.NoSerde{}}
2024-11-26 15:32:07 [ℹ]  will use version 1.29 for new nodegroup(s) based on control plane version
Error: no output "VPC" in stack "eksctl-prod-1-cluster-addon-iamserviceaccount-kube-system-aws-load-balancer-controller-prod-1-cluster"

Versions

$ eksctl info
eksctl version: 0.185.0
kubectl version: v1.30.0-eks-036c24b
OS: linux

[github-actions]

Hello NicolasGrosjeanProbayes 👋 Thank you for opening an issue in eksctl project. The team will review the issue and aim to respond within 1-5 business days. Meanwhile, please read about the Contribution and Code of Conduct guidelines here. You can find out more information about eksctl on our website

[thurahtetaung]

I am having the same problem. Looks like eksctl is mistaking a service account cloudformation stack as the cluster stack and looking for VPC info in that stack. My solution in our dev cluster was to delete the service account mentioned in the error and then it works but after I created the service account again, the issue reoccurs. Unfortunately, this cannot be used to solve in our prod cluster as we cannot delete the service account. Is there anyway to point the eksctl to the right cluster stack to solve it?