ebs-csi-node 守护进程集未在 arm64 节点上运行

Question

I am using eks cluster with a mix of amd64 and arm64 nodes, after deploying the aws-ebs-csi driver with helm chart 2.3.0 to my cluster found the ebs-csi-node daemon set is only created on amd64 node not on arm64 due which volume is not getting provisioned.

NAME           DESIRED   CURRENT   READY   UP-TO-DATE   AVAILABLE   NODE SELECTOR            AGE
aws-node       192       192       192     192          192         <none>                   691d
ebs-csi-node   19        19        19      19           19          kubernetes.io/os=linux   256d
kube-proxy     192       192       192     192          192         <none>                   691d

Also running 2 replicas of ebs-csi-controller both of them are getting deployed only on amd64 node. Do I need to any selector to helm the chart to choose the node on which ebs-csi-controller need to deploy?

Answer 1

I am using the terraform resources to configure the ebs-csi-controller in the eks cluster as shown:

resource "helm_release" "aws-ebs-csi-driver" {
  name       = "aws-ebs-csi-driver"
  repository = "https://kubernetes-sigs.github.io/aws-ebs-csi-driver"
  chart      = "aws-ebs-csi-driver"
  namespace  = "kube-system"
  version = "2.3.0"

  values = [
    <<EOF
enableVolumeScheduling: true
enableVolumeResizing: true
enableVolumeSnapshot: true
node.tolerateAllTaints: true
EOF
  ]
}

Found that even after setting the "node.tolerateAllTaints: true" value the DS is not getting updated with the torelation "- operation: exists" after manually updating the torelation DS got created on all nodes. I am in correcting set the torelation in terraform. After updating the value like this :

values = [
    <<EOF
enableVolumeScheduling: true
enableVolumeResizing: true
enableVolumeSnapshot: true
node:
  tolerateAllTaints: true
EOF
  ]

DS got created on all the node