NFS connection timing out on EKS

10/15/2019

I have an NFS helm chart. It is one of the charts for an application that has 5 more sub-charts. 2 of the charts have a shared storage which I am using NFS. In GCP when I provide NFS service name in the PV it works.

apiVersion: v1
kind: PersistentVolume
metadata:
  name: {{ include "nfs.name" . }}
spec:
  capacity:
    storage: {{ .Values.persistence.nfsVolumes.size }}
  accessModes:
    - {{ .Values.persistence.nfsVolumes.accessModes }}
  mountOptions:
    - nfsvers=4.1
  nfs:
    server: nfs.default.svc.cluster.local # nfs is from svc {{ include "nfs.name" .}}
    path: "/opt/shared-shibboleth-idp"

But the same doesn't work on AWS EKS. The error there - on AWS EKS - is connection timeout so it can't mount the volume. When I change the server to

server: a4eab2d4aef2311e9a2880227e884517-1524131093.us-west-2.elb.amazonaws.com .

I get connection timed out.

All the mounts are okay since it works well with GCP. What am I doing wrong?

-- Shammir
amazon-eks
amazon-web-services
kubernetes
kubernetes-helm
nfs

0 Answers