README.md

Step 6

Play with nfs and volumes

In order to make it easy, lets create a nfs server on the master. This is thought to be a test scenario please read more documentation about how to make it secure if you want to run in real environment

sudo apt-get install nfs-kernel-server rpcbind -y
sudo mkdir -p /export //create a directory to share
sudo chmod -R 777 /export 
sudo vim /etc/exports (Add this line: /export       172.17.4.0/24(rw,fsid=0,insecure,no_subtree_check,async)
sudo vim /etc/hosts.allow (Add this line: rpcbind mountd nfsd statd lockd rquotad : ALL)
sudo service nfs-kernel-server restart

And voila, we have a nfs server (open to the world) to mount in our pods Now lets add the config file here to be read for the app

echo "title=\"My nfs pv works\"" > /export/config.ini

In the nodes now we should install the client

sudo apt-get install rpcbind nfs-common
sudo vim /etc/hosts.allow (Add this line: rpcbind : ALL)

Next step is add the PV with the server IP and the path, so kubernetes will prepare it to be requested.

kubectl apply -f nfs-pv.yaml
kubectl get pv 

We should it has the correct status Now we create a volume claim so the pod can request this storage

kubectl apply -f nfs-pvc.yaml
kubectl get pvc 

Here k8s should display the status bound after some seconds. Then we are ready to create our pod with a nfs volume

Here we has change the spec to have a nfs volume with the correct pvc

cat single_pod.yaml

Last step, is run it and check if our pod has mounted and we see the correct configuration

kubectl apply -f single_pod_basic.yaml
kubectl describe pod single-pod
kubectl exec single-pod -- curl localhost:8080

And you should see the correct message. Now this pod can be schedule to the other node or replicated and the volume will last