Skip to main content

AWS EFS CSI Driver

Wrapper chart for the upstream aws-efs-csi-driver chart (pinned to 4.4.1, appVersion 3.4.1), the CSI driver that provisions AWS EFS (NFS-backed, ReadWriteMany) volumes as PersistentVolumes.

Why it's in KubeAid​

Not currently rendered by kubeaid-cli's bootstrap flow — there is no AWS*TemplateNames entry for it in pkg/constants/templates.go. It's an opt-in chart for AWS clusters that need shared/ReadWriteMany storage (EBS is block storage, single-attach); enable it manually as an Argo CD Application the same way other argocd-helm-charts/ charts are wired in.

Prerequisites​

  • An EFS filesystem and access point(s) provisioned out-of-band (Terraform/Crossplane/console).
  • IAM permissions for the driver identity to call the EFS API (mount targets, access points).

Key values / KubeAid-specific configuration​

Upstream values live under the aws-efs-csi-driver: key.

  • replicaCount: 1 (values.yaml) — single controller replica.
  • networkpolicies (default false, not present in values.yaml but read by the template): when set to true, templates/netpol-aws-efs-csi-driver.yaml renders a Calico-flavored crd.projectcalico.org/v1 NetworkPolicy restricting egress to the apiserver and to kube2iam (port 8181) — a legacy IAM-via-annotations pattern that predates IRSA; treat it as historical unless you're still running kube2iam.

Operational notes​

  • docs/guides/release.md explicitly excludes aws-efs-csi-driver from manage-helm-chart.sh --update-all, so it does not track upstream automatically the way most other vendored charts do — bump it manually.
  • No StorageClass is shipped by this wrapper; create one (with a fileSystemId parameter, typically provisioningMode: efs-ap for per-PVC access points) per cluster.