Glusterfs is an open source scale-out filesystem. These examples provide information about how to allow containers use Glusterfs volumes.

The example assumes that the Glusterfs client package is installed on all nodes.

Install Glusterfs client package on the Kubernetes hosts.

The following volume spec illustrates a sample configuration.

{
     "name": "glusterfsvol",
     "glusterfs": {
        "endpoints": "glusterfs-cluster",
        "path": "kube_vol",
        "readOnly": true
    }
}

The parameters are explained as the followings.

• endpoints is endpoints name that represents a Gluster cluster configuration. kubelet is optimized to avoid mount storm, it will randomly pick one from the endpoints to mount. If this host is unresponsive, the next Gluster host in the endpoints is automatically selected.
• path is the Glusterfs volume name.
• readOnly is the boolean that sets the mountpoint readOnly or readWrite.

Detailed POD and Gluster cluster endpoints examples can be found at v1beta3/ and endpoints/

# create gluster cluster endpoints
$ kubectl create -f examples/glusterfs/endpoints/glusterfs-endpoints.json
# create a container using gluster volume
$ kubectl create -f examples/glusterfs/v1beta3/glusterfs.json

Once that's up you can list the pods and endpoint in the cluster, to verify that the master is running:

$ kubectl get endpoints
$ kubectl get pods

If you ssh to that machine, you can run docker ps to see the actual pod and mount to see if the Glusterfs volume is mounted.