Fc2 Vfs kernel panic after upgrade from Fc1 using Volume Management

[Alexandre Oliva]

On May 20, 2004, Peter Smith <peter.smith at utsouthwestern.edu> wrote:

> Originally I had installed Fc1 and opted for the Volume Management
> arrangement.  The default was to setup root (/) as
> /dev/Volume00/LogVol01 .  After upgrading to Fc2,

How did you upgrade?  Anaconda, or up2date, yum or something else that
couldn't possibly have fixed things up?

I don't know whether Anaconda does anything to fix LVM upgrades, but I
vaguely remember some note about this not working in FC2 test releases
Release Notes, and I can't find it in the FC2 final release notes, so
I assume it should work.

Anyhow, if it doesn't, or if you chose one of the other upgrade paths,
the way to fix it is to remove the LVM1 device nodes from the VG
directories.  After that, a reboot should fix things up.  If it
doesn't, run as root, after entering the password when the system
fails to boot up:

mount -o remount,rw /
echo mkdmnod | nash
vgscan
vgchange -ay
vgmknodes

this was required in early FC2 test releases, when initrd attempted to
use device nodes from the root filesystem, instead of from initrd.
I don't think this should still be necessary, but if everything else
fails, it's worth a shot.


After setting things up, you may want to re-create initrd, just in
case.

-- 
Alexandre Oliva             http://www.ic.unicamp.br/~oliva/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}