Hi!
I have the following problem:
kernel 2.4.12, md and raid1 compiled into the kernel.
/dev/hdb old linux installation
/dev/md0 -> /dev/hde1,/dev/hdg1 new installation
When I boot my old installation the md device is automatically configured
by the kernel and I can mount it (reiserfs) without any problems.
When I try to boot the new installation with the same kernel the md device
is initialized, but the kernel cannot mount the root device. I get msgs
about FAT problems and about mounting root as msdos.
Here some config files:
lilo.conf:
image = /boot/lx-2.4.12
root = /dev/hdb1
label = old
image = /boot/lx-2.4.12
root = /dev/md0
label = new
optional
Thanks a lot for any information
Best wishes
Norbert
-----------------------------------------------------------------------
Norbert Preining <[email protected]>
University of Technology Vienna, Austria gpg DSA: 0x09C5B094
-----------------------------------------------------------------------
LOUTH (n.)
The sort of man who wears loud check jackets, has a personalised
tankard behind the bar and always gets served before you do.
--- Douglas Adams, The Meaning of Liff
On Thu, 18 Oct 2001, Norbert Preining wrote:
> Hi!
>
> I have the following problem:
>
> kernel 2.4.12, md and raid1 compiled into the kernel.
> /dev/hdb old linux installation
> /dev/md0 -> /dev/hde1,/dev/hdg1 new installation
>
> When I boot my old installation the md device is automatically configured
> by the kernel and I can mount it (reiserfs) without any problems.
>
> When I try to boot the new installation with the same kernel the md device
> is initialized, but the kernel cannot mount the root device. I get msgs
> about FAT problems and about mounting root as msdos.
>
> Here some config files:
> lilo.conf:
> image = /boot/lx-2.4.12
> root = /dev/hdb1
> label = old
> image = /boot/lx-2.4.12
> root = /dev/md0
> label = new
> optional
To use a md as root, you need to add a kernel command line:
md0=1,/dev/hde1,/dev/hdg1
Put that in the append= line of lilo.conf or type it at the lilo command
prompt.
See also Documentation/md.txt in the Linux source tree.
-jwb
Norbert Preining <[email protected]> :
[...]
> When I try to boot the new installation with the same kernel the md device
> is initialized, but the kernel cannot mount the root device. I get msgs
> about FAT problems and about mounting root as msdos.
$ less ChangeLog
[pre1]
- Al Viro: fix partition handling sanity check.
Can you reproduce it with >= 2.4.13-pre1 ?
We've got the same kind of lilo.conf (no md=... option) and it works
here for vanilla to 2.4.10 and -ac to 2.4.12-ac3 (no vanilla 2.4.12 tested
in a raid setup, surprise !). I would test 2.4.13-prex before modifying
lilo.conf.
--
Ueimor
"Jeffrey W. Baker" wrote:
>
> On Thu, 18 Oct 2001, Norbert Preining wrote:
>
> > Hi!
> >
> > I have the following problem:
> >
> > kernel 2.4.12, md and raid1 compiled into the kernel.
> > /dev/hdb old linux installation
> > /dev/md0 -> /dev/hde1,/dev/hdg1 new installation
> >
> > When I boot my old installation the md device is automatically configured
> > by the kernel and I can mount it (reiserfs) without any problems.
> >
> > When I try to boot the new installation with the same kernel the md device
> > is initialized, but the kernel cannot mount the root device. I get msgs
> > about FAT problems and about mounting root as msdos.
> >
> > Here some config files:
> > lilo.conf:
> > image = /boot/lx-2.4.12
> > root = /dev/hdb1
> > label = old
> > image = /boot/lx-2.4.12
> > root = /dev/md0
> > label = new
> > optional
>
> To use a md as root, you need to add a kernel command line:
>
> md0=1,/dev/hde1,/dev/hdg1
>
> Put that in the append= line of lilo.conf or type it at the lilo command
> prompt.
>
> See also Documentation/md.txt in the Linux source tree.
The line you provide doesn't look anything like the two forms in the
md.txt you mention. Or rather it looks like a blending, but neither of
them is md0= in form. I have to look at the code to see which is
correct, possibly yours, since the
append = "md=0,/dev/sda1,/dev/sdb1"
line doesn't seem to work :-(
The md.txt says the persistent superblock form is:
md=<md device no.>,dev0,dev1,...,devn
which doesn't seem to work for me.
--
bill davidsen ([email protected])
Prodigy Internet Server Group
Project Leader, USENET news
914-448-1241
On Tuesday October 23, [email protected] wrote:
>
> The line you provide doesn't look anything like the two forms in the
> md.txt you mention. Or rather it looks like a blending, but neither of
> them is md0= in form. I have to look at the code to see which is
> correct, possibly yours, since the
> append = "md=0,/dev/sda1,/dev/sdb1"
> line doesn't seem to work :-(
Odd ... I use lines just like that. e.g.:
append="md=0,/dev/hda1,/dev/hde1,/dev/hdg1"
and it works just fine. What do you get in the way of error messages?
NeilBrown
In article <[email protected]>,
Neil Brown <[email protected]> wrote:
| On Tuesday October 23, [email protected] wrote:
| >
| > The line you provide doesn't look anything like the two forms in the
| > md.txt you mention. Or rather it looks like a blending, but neither of
| > them is md0= in form. I have to look at the code to see which is
| > correct, possibly yours, since the
| > append = "md=0,/dev/sda1,/dev/sdb1"
| > line doesn't seem to work :-(
|
| Odd ... I use lines just like that. e.g.:
| append="md=0,/dev/hda1,/dev/hde1,/dev/hdg1"
|
| and it works just fine. What do you get in the way of error messages?
None - the system simply exits the BIOS, reads the first drive once and
cold boots. The drive is okay, I can read both copies of the mirror end
to end without error after booting from floppy. Lilo claims it writes to
the md0 device, but boot fails.
--
bill davidsen <[email protected]>
His first management concern is not solving the problem, but covering
his ass. If he lived in the middle ages he'd wear his codpiece backward.
> In article <[email protected]>,
> Neil Brown <[email protected]> wrote:
> | On Tuesday October 23, [email protected] wrote:
> | >
> | > The line you provide doesn't look anything like the two forms in the
> | > md.txt you mention. Or rather it looks like a blending, but neither of
> | > them is md0= in form. I have to look at the code to see which is
> | > correct, possibly yours, since the
> | > append = "md=0,/dev/sda1,/dev/sdb1"
> | > line doesn't seem to work :-(
> |
> | Odd ... I use lines just like that. e.g.:
> | append="md=0,/dev/hda1,/dev/hde1,/dev/hdg1"
> |
> | and it works just fine. What do you get in the way of error messages?
>
> None - the system simply exits the BIOS, reads the first drive once and
> cold boots. The drive is okay, I can read both copies of the mirror end
> to end without error after booting from floppy. Lilo claims it writes to
> the md0 device, but boot fails.
I always let lilo write to the first and the second disk itself, and I
do not use any kernel parameters for sw raid. Is writing to the disks
instead of writing to /dev/md0 the wrong way? It works for me.
Ookhoi
On Thu, 25 Oct 2001, Ookhoi wrote:
> > In article <[email protected]>,
> > Neil Brown <[email protected]> wrote:
> > | On Tuesday October 23, [email protected] wrote:
> > | Odd ... I use lines just like that. e.g.:
> > | append="md=0,/dev/hda1,/dev/hde1,/dev/hdg1"
> > |
> > | and it works just fine. What do you get in the way of error messages?
> >
> > None - the system simply exits the BIOS, reads the first drive once and
> > cold boots. The drive is okay, I can read both copies of the mirror end
> > to end without error after booting from floppy. Lilo claims it writes to
> > the md0 device, but boot fails.
>
> I always let lilo write to the first and the second disk itself, and I
> do not use any kernel parameters for sw raid. Is writing to the disks
> instead of writing to /dev/md0 the wrong way? It works for me.
Tried that as well, with -b, still doesn't like to boot. It may well be
the system, which is seriously strange in config. I have some other
systems I'd love to install RAID, but the Redhat (mandated by management)
install won't do RAID unless graphical install. Unless that's fixed/added
in 7.2 which I haven't tried. These systems don't do graphical...
--
bill davidsen <[email protected]>
CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.