Subject: Re: wedges and what does that mean?
To: Thor Lancelot Simon <email@example.com>
From: Bill Studenmund <firstname.lastname@example.org>
Date: 09/07/2006 11:22:44
Content-Type: text/plain; charset=us-ascii
On Thu, Sep 07, 2006 at 01:46:18PM -0400, Thor Lancelot Simon wrote:
> On Thu, Sep 07, 2006 at 10:43:28AM -0700, Bill Studenmund wrote:
> > On Wed, Sep 06, 2006 at 09:58:54AM +0200, Pavel Cahyna wrote:
> > > On Tue, Sep 05, 2006 at 10:31:08AM -0700, Bill Studenmund wrote:
> > > > The up-shot is that we need a way to rename and re-minor-number dev=
> > > > Combine that with a daemon that keeps state in userland, and we for=
> > > > wedges to come back with the same minor # as before.
> > >=20
> > > The point is, we don't have such thing yet. Shouldn't we delay migrat=
> > > to wedges?
> > No, as we'll never get it if we keep delaying.
> I cannot believe you seriously mean to suggest that it is worth breaking
> operation as NFS server in a severe way to avoid further delay migrating
> to wedges. Is that really what you mean?
It only breaks things (now) if you don't wire things down. Scan for wedges=
in a script so the order stays the same, and you'll get consistent wedge=20
Also, since we haven't turned off the old partition method (for disks=20
partitioned where it works), mount your file systems off of them for now=20
if you're concerned.
I agree this needs fixing, but you are bringing a fervor to this=20
discussion that seems excessive. We have until 5.0 to fix this, and I'm=20
sure we will. And if we don't, we definitely won't turn off the old=20
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (NetBSD)
-----END PGP SIGNATURE-----