As explained in the commit summary, if VDSM is going to create
metadata and image directories, then virt-v2v should not mkdir those
same directories (and hence fail).
I'm not really convinced this is the right approach.
I think instead of adding more and more hacks on top of '-o rhev'
which should split this output mode into two sub modes, one for
outputting to the RHEV Export Storage Domain (ie. "traditional"
imports) and one which can handle the --rhev* parameters and the Data
domain ('-o vdsm' perhaps?)
Rich.