K
Q

Is NFS hard- or soft-mounted to Kubernetes Pods?

9/22/2015

Is NFS hard- or soft-mounted when making a Pod in Kubernetes with an NFS-volume?

As I understand this might have an impact on how it handles a timeout?

Example yaml:

apiVersion: v1
kind: Pod
metadata:
  name: nfs-web
spec:
  containers:
- name: web
  image: nginx
  ports:
    - name: web
      containerPort: 80
  volumeMounts:
      - name: nfs
        mountPath: "/usr/share/nginx/html"
  volumes:
    - name: nfs
      nfs:
        server: nfs-server.default.kube.local
        path: "/"
-- NegatioN
kubernetes

Similar Questions

kublet doesn't start with cloud provider "openstack"
Kubernetes whitepaper
How to run an etcd cluster among pod replicas?
pod is not showing in ready state

1 Answer

9/22/2015

I believe that the NFS mounts inside of a POD will use the defaults provided by the implementation of NFS in the container OS. I can't be 100% certain (I'm not deeply familiar with the code), but in my experience the NFS mounts are mounted with the hard option, which is default in most implementations of NFS (see man nfs for more details on your OS; soft is often considered dangerous.)

The NFSVolumeSource struct doesn't appear to have the ability to know about mount settings (except read-only) and I don't see any hard-coded options in the NFS volume code.

You can check on your own PODs with something like this to gather the NFS options in use:

$ kubectl exec nfs-web-<XXXXX> -c web -- mount|grep nfs

-- rwehner
Source: StackOverflow