Description of problem:
When a volume creation fails, extended attributes are left behind on any bricks that were not part of that failure.

Version-Release number of selected component (if applicable):
3.6.2

How reproducible:
Always

Steps to Reproduce:
1. Pick a brick and do something to ensure it fails, make the filesystem read only, use selinux to block access, whatever. Call this server3:/brick1
2. Create a volume: gluster volume create fubar server{1,2,3}:/brick1
3. When volume creation fails on server[12], getfattr -n trusted.glusterfs.volume-id -e hex /brick1

Actual results:
volume-id exists preventing additional attempts at creating a volume with those server/brick combinations.

Expected results:
A failed volume creation should return the potential brick to pre-attempt state.

Additional info:
This is a common problem seen in #gluster. Users do not understand, based on the error message, what the cause of the problem is nor how to fix it.