[Bug] eksctl create Capacity Reservation for ML will lost CapacityReservationTarget data

[xzp1990]

What were you trying to accomplish?

use eksctl to create a managed nodegroup with Capacity reservation for ML (p5.48xlarge and p5en.48xlarge)

What happened?

Create nodegroup fail and the checking the cloudformation and find the CapacityReservationTarget is empty:
"CapacityReservationSpecification": {
"CapacityReservationTarget": {}
},
"InstanceMarketOptions": {
"MarketType": "capacity-block"
},

How to reproduce it?

apiVersion: eksctl.io/v1alpha5
kind: ClusterConfig

metadata:
name: aws-us-east-01
region: us-east-1

managedNodeGroups:

• name: p5-ng-efa
  instanceType: p5.48xlarge Add version command #8 x 3.84 TB NVMe SSD
  capacityReservation:
  capacityReservationTarget:
  capacityReservationId: "cr-0a91582543292e10a" # Replace with your ML Capacity Block ID
  instanceMarketOptions:
  marketType: capacity-block
  privateNetworking: true
  efaEnabled: true
  iam:
  attachPolicyARNs:
  • arn:aws:iam::aws:policy/AmazonEKSWorkerNodePolicy
  • arn:aws:iam::aws:policy/AmazonEKS_CNI_Policy
  • arn:aws:iam::aws:policy/AmazonEC2ContainerRegistryReadOnly
  • arn:aws:iam::aws:policy/AmazonSSMManagedInstanceCore
  • arn:aws:iam::aws:policy/CloudWatchAgentServerPolicy
    desiredCapacity: 1
    minSize: 1
    maxSize: 1
    volumeSize: 200
    tags:
    "eks/node-type": "gpu"
    labels:
    eks/node-type: gpu
    eks/gpu-type: nvidia
    nvidia.com/gpu.present: "true"
    taints:
  • key: nvidia.com/gpu
    effect: "NoSchedule"
    subnets:
    • subnet-01290a2d94274f6ee #AZ E, IAD5
      ssh: # use existing EC2 key
      publicKeyName: LabKey

Logs
part of the yaml:
"LaunchTemplate": {
"Type": "AWS::EC2::LaunchTemplate",
"Properties": {
"LaunchTemplateData": {
"BlockDeviceMappings": [
{
"DeviceName": "/dev/xvda",
"Ebs": {
"Iops": 3000,
"Throughput": 125,
"VolumeSize": 200,
"VolumeType": "gp3"
}
}
],
"CapacityReservationSpecification": {
"CapacityReservationTarget": {}
},
"InstanceMarketOptions": {
"MarketType": "capacity-block"
},
"InstanceType": "p5.48xlarge",
"KeyName": "LabKey",
"MetadataOptions": {
"HttpPutResponseHopLimit": 2,
"HttpTokens": "required"
},

Anything else we need to know?
N/A
It's very easy to reproduce.

Versions
Admin:~/environment/koala $ eksctl info
eksctl version: 0.215.0
kubectl version: v1.31.0-eks-a737599
OS: linux

[NicholasBlaskey]

Hello,

Thanks for the bug report. I was able to reproduce this, I'm going to work on putting up a fix for this soon.

[NicholasBlaskey]

@xzp1990 on second look I noticed a typo in your Config file

Can you try changing Id to ID

- capacityReservationId: "cr-0a91582543292e10a" # Replace with your ML Capacity Block ID
+ capacityReservationID: "cr-0a91582543292e10a" # Replace with your ML Capacity Block ID

[NicholasBlaskey]

We aren't getting you an invalid error message due to #753

and
kubernetes-sigs/yaml#15

[xzp1990]

@NicholasBlaskey Nick, thanks for you help, it works now:
from Cloudformation, I can see the CapacityReservationId is not empty.
"CapacityReservationSpecification": {
"CapacityReservationTarget": {
"CapacityReservationId": "cr-0a91582543292e10a"
}
},

[NicholasBlaskey]

Thanks! Sorry about the not clear error message