[libvirt] libvirt and

Mark McLoughlin markmc at redhat.com
Thu Dec 18 08:21:01 UTC 2008

On Wed, 2008-12-17 at 22:53 -0800, H. Peter Anvin wrote:

> I'm kind of wondering why libvirt defaults to by
> default.

Randomly chosen out of pure ignorance :-)


>   It is a piece of address space which is relatively likely to
> conflict with address space used in the environment surrounding the
> machine.  Since libvirtd is on by default if installed, this is
> particularly problematic.
> I would like to suggest either one of the following as default address
> space, unless the user has explicitly configured otherwise:
>	- reserved as "test and example network"
>	- reserved as "benchmark test network"
> See RFC 3330 and RFC 2544 for the definitions of these networks.  Both
> of them are "should not appear on the public Internet" address blocks,
> and much less likely than the RFC 1918 address block (,
>, and to be encountered "in the wild" by a
> user.

Neither network sounds like it exactly fits our needs, but we're
probably better off with than with what we have now. We're
ignoring the fact that it's "for use in documentation and example code"
and just relying on the fact that it should not be seen in the wild.

Thanks for bringing it up ...


More information about the libvir-list mailing list