Description of problem:
Since Dec23, packstack --allinone failed with:

192.168.0.164_cinder.pp:                          [ [0;31mERROR[0m ]
Applying Puppet manifests                         [ [0;31mERROR[0m ]

[0;31mERROR : Error appeared during Puppet run: 192.168.0.164_cinder.pp
Error: cinder type-create iscsi returned 1 instead of one of [0][0m
You will find full trace in log /var/tmp/packstack/20151223-165310-Lw1D16/manifests/192.168.0.164_cinder.pp.log

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


How reproducible:

The RDO release: https://repos.fedorapeople.org/repos/openstack/openstack-kilo/rdo-release-kilo-1.noarch.rpm
The Packstack command line: sudo packstack --allinone --os-swift-install=y --os-ceilometer-install=n --nagios-install=n --provision-demo=n --keystone-admin-passwd=sf4ever --os-heat-install=y --cinder-volumes-size=50G

Steps to Reproduce:
You can currently observe that bug in this job: http://softwarefactory-project.io/jenkins/job/sf-openstack-kilo-1node-allinone/
starting with build #182

Actual results:
Packstack fail to deploy

Expected results:
Packstack succeed to deploy

Additional info:
We hit that bug in software factory project CI (thus the sfci Keywords).
Two jobs are testing OpenStack integration with the SF project using RDO release and packstack deployment:
  sf-openstack-kilo-1node-allinone and sf-openstack-liberty-1node-allinone.
The two jobs are identique, except that one will use rdo release, the other kilo release.