I still have to same problem that newer kernels from Debian do not
recognize a LVM during boot. I already send this problem to this mailing
list 7 weeks ago (see below) but did not get any answer.
Today, I wanted to upgrade a etch'n'half kernel from
got the same result. So, the kernel or the configuration is not the
problem (/boot/condfig) is the same. It must be something different. It
looks that it depends when the kernel has been compiled or installed...
I updated a server from Etch to Lenny. It uses a external storage
solution to store a lot of data that is connected via SCSI. The server
has a PCI card that is used by the sym53c8xx kernel module. On the
storage solution are several physical disk that are combined by LVM on
But when I rebooted the server with now Lennys 2.6.26er kernel and the
userspace apps the LVM volume was not regonized.
However, if I ignored the errors that the volume could not be mounted
and used the login and did the following
/sbin/vgchange -a y
the volume was recognized and useable. I wondered why this worked and
had a closer look at the startup script of lvm in /etc/init.d/ but they
looked sufficient. I also put the two commands in an additional script
that is run after the normal lvm scripts and before S30checkfs.sh and
S35mountall.sh but it also did not worked.
If I boot a kernel from Etch the volume is recognized.
So I guess that it is a bug of the kernel or the packaging... What do
Dipl.-Wirtsch.-Ing. Martin Bretschneider
Institut für Mess- und Regelungstechnik
Leibniz Universität Hannover
Nienburger Straße 17
Telefon: +49 511 762-4456
Fax: +49 511 762-3234
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact firstname.lastname@example.org