We've had a chat here at L'pool, and we're of the opinion that it would be less of an upheaval to always roll the systems forward, in the absence of a robust mechanism to automatically ensure the integrity of a rollback. That would imply taking the last working rpm, and re-stamping the version number, as suggested by others. We understand there may be difficulties achieving this, but we're not looking forward to fishing out changes one by one if the site is down. Steve -- Steve Jones [log in to unmask] Grid Administrator office: 220 High Energy Physics Division tel (int): 42334 Oliver Lodge Laboratory tel (ext): +44 (0)151 794 2334 University of Liverpool http://www.liv.ac.uk/physics/hep/