On Tue, Aug 23, 2011 at 08:12:46AM -0600, Eric Blake wrote: > On 08/23/2011 05:18 AM, Daniel P. Berrange wrote: > >> <entry name='uuid'>00010203-0405-0607-0809-0A0B0C0D0E0F</entry> > > >Yes, your hardware vendor is being very very bad :-( > > >So, libvirtd provides a way to override the default host UUID, > >by setting 'host_uuid' in /etc/libvirt/libvirtd.conf, which > >should make the migration sanity check pass. > > Should we also teach libvirt to recognize this particular UUID as > invalid, just as we already reject all 0's and all F's as bad (also > values seen from bad hardware vendors)? Once we start going down this route, it'll never end IMHO. I've seen so many "bad" patterns used by various vendors... The alternate is to not use the SMBIOS uuid, but rather rely on the fact that new hosts with SystemD, have a sane UUID generated and recorded in /etc/machine-id Daniel -- |: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :| |: http://libvirt.org -o- http://virt-manager.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
