Description of problem:
=======================
In a 1 x 2 replicate volume a server node crashed. The node comes online. Even before mounting bricks to a valid device containing xfs if we perform "gluster volume start <volume_name> force" we assign "trusted.glusterfs.volume-id" to the brick without checking whether the brick is not a mount point with xfs. 

This results in unsupported configuration (no xfs/LVM) and worse may fill up / of the storage node if not discovered.

Refer to https://bugzilla.redhat.com/show_bug.cgi?id=860999 for the case which filled "/" file system. 

Version-Release number of selected component (if applicable):
=============================================================
glusterfs 3.4.0.43.1u2rhs built on Nov 12 2013 07:38:20

How reproducible:
=================
Often

Steps to Reproduce:
===================
1. Create a 1 x 2 replicate volume with brick mounted on a mount point with xfs. Do not add the automount of these bricks in "/etc/fstab"

2. Restart the node2.

3. When node2 comes online execute: "gluster v start <volume_name> force"

Actual results:
================
Assigns the "trusted-volume-id" to the brick and starts the brick process. 

Expected results:
================
"gluster v start <volume_name> force" should check if the specified bricks are separate mount points with xfs on LVM volumes and reject the command or at least issue a warning