I don't think that having different partitioning schemes is a
problem for an LVM. Whatever the partitioning scheme, you would
simply have a single "NetBSD LVM" partition which goes into the
pool with the LVM partitions on the other disks. The traditional
disklabel then goes away (or is constructed automatically to reflect
the partitioning scheme, but not used for the LVM pieces) to be
replaced with a more complicated but more flexible LVM scheme built
on top of "NetBSD LVM" partitions.

					Gary D. Duzan



In Message <Pine.NEB.4.21.0006131136210.6256-100000@skullport.z.zembu.com> ,
   Bill Studenmund <wrstuden@zembu.com> wrote:

=>On Tue, 13 Jun 2000, David Rankin wrote:
=>
=>> The "real solution" would be to have a volume manager (similar to IBM's
=>> LVM, Vinum, etc.), rather than trying to hack the partitioning scheme.
=>> Is there anyone out there working on such a beast? (Before you ask, I have
=>> been contemplating such a project, but my employer and I are currently
=>> discussing whether my off-hours work on one would conflict with my 
=>> employment agreement (*blah*).)
=>
=>No, actually, it's not.
=>
=>While LVM is cool, and I'd _love_ it if you could work on it, it's not the
=>"real solution." First off, we have lots of disks we share with other
=>OS's. We need to share partitioning with them. While you could hack that
=>into an LVM system, it'd be a hack.
=>
=>Part of the goal of what we (I) want to do when we go to 64 partitions is
=>gain the ability to grok multiple partitioning schemes. So say we could
=>find all the NetBSD and FreeBSD or MacOS partitions on a disk. I think
=>that is a contrary thing to how LVM's work (as I understand them).
=>
=>I hope your negotiations work out!
=>
=>Take care,
=>
=>Bill
=>