After a crash test, we unplug the power of the primary master node, the FOSS-Cloud was still stable, vm's, former running on the primary master node, could be migrated to the other node and just restarted. The golden image also pointed to the primary master node but could not be migrated to the running node. We tried to deselect the primary master node in the pool without success, because of supposed running dynamic vm's. Therefore, we were forced to delete the golden image and to recopy it from the template.