boot failure "mount: mounting /dev/sda1 on /root failed: no such device"
Kelly Clowers
I get error messages like this:
mount: mounting /dev/sda1 on /root failed: no such device
mount: mounting /sys on /root/sys failed: no such file or directory
mount: mounting /proc on /root/proc failed: no such file or directory
Target filesystem doesn't have /sbin/init
/bin/sh: can't access tty: job control turned off
I have tried different kernels (.29 and .30 - Sid packages), I have
chrooted from a LiveCD and updated/reinstalled a number of
packages: initramfs-tools, udev, and various sysvrc and init
packages. I have regenerated the initramfs. I double-checked
fstab, and I have tried using root=/dev/sda1 and root=UUID=
Nothing has worked, any further ideas are welcome.
Thanks,
Kelly Clowers
--
To UNSUBSCRIBE, email to debian-us...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listm...@lists.debian.org
Nuno Magalhães
I had this issue after upgrading from 2.6.29-1-amd64 to
2.6.29-2-amd64, all i did was reboot using the -1-. Upgraded again and
now i'm using 2.6.30-1-amd64 since the previous is also failing. If
it's the same problem it's not HD-related.
HTH
--
() ascii ribbon campaign - against html e-mail
/\ ascii-rubanda kampajno - kontraŭ html-a retpoŝto
Kelly Clowers
>> Nothing has worked, any further ideas are welcome.
>
> I had this issue after upgrading from 2.6.29-1-amd64 to
> 2.6.29-2-amd64, all i did was reboot using the -1-. Upgraded again and
> now i'm using 2.6.30-1-amd64 since the previous is also failing. If
> it's the same problem it's not HD-related.
I didn't really think it was HD-related, as sda1 mounts fine
on a LiveCD, and fsck.ext4 finds no problems.
As for the kernel version, I have tried 2.6.30-1-amd64,
and I get the same result. It is really baffling to me.
Is there a way to get a shell inside the initramfs, to
look around?
Thanks,
Kelly Clowers
Kelly Clowers
> 2009/7/7 Nuno Magalhães <nunoma...@eu.ipp.pt>:
>>> Nothing has worked, any further ideas are welcome.
>>
>> I had this issue after upgrading from 2.6.29-1-amd64 to
>> 2.6.29-2-amd64, all i did was reboot using the -1-. Upgraded again and
>> now i'm using 2.6.30-1-amd64 since the previous is also failing. If
>> it's the same problem it's not HD-related.
>
> I didn't really think it was HD-related, as sda1 mounts fine
> on a LiveCD, and fsck.ext4 finds no problems.
>
> As for the kernel version, I have tried 2.6.30-1-amd64,
> and I get the same result. It is really baffling to me.
>
> Is there a way to get a shell inside the initramfs, to
> look around?
Anyone have any other ideas about this? Because I don't
have any new ideas. Nothing I have tried works and I am
completely stumped. I just tried the most recent kernel
update, etc. and the result was the same.
Raphael Bossek
Hi,
I'm a little bit happy reading this because my Debian system fails form one day to another (after an upgrade of the linux kernel 2.6.26-2-686). I've started a discussion about my broblem already without any solution.
It would be interesting for me to know if you partitions are deteced stering the driver. Please grep dmsg for your device node and check if they are within /proc/partitions too. If not, try to call hdparm -z /dev/what-ever-your-device-ist to force the kernel to reload the partition table.
To be able to start hdparm you have to copy the app from somewhere else because it's not part of initrd environment. I used an external USB HD witch is detected by udev seamless.
My next try will be to go back to an older linux version and udev. Maybe udev is also an problem.
My problem is "Resourse or device busy" at this moment within the initrd environment afer hparm -z reloaded the partitiontable and udev created the device node.
Regards,
Raphael
On Jul 13, 2009 7:31 PM, "Kelly Clowers" <kelly....@gmail.com> wrote:On Tue, Jul 7, 2009 at 12:26, Kelly Clowers<kelly....@gmail.com> wrote: > 2009/7/7 Nuno Magalhãe...
Anyone have any other ideas about this? Because I don't
have any new ideas. Nothing I have tried works and I am
completely stumped. I just tried the most recent kernel
update, etc. and the result was the same.
Thanks, Kelly Clowers -- To UNSUBSCRIBE, email to debian-us...@lists.debian.org with a ...
Kelly Clowers
Bossek<raphael...@googlemail.com> wrote:
> Hi,
>
> I'm a little bit happy reading this because my Debian system fails form one
> day to another (after an upgrade of the linux kernel 2.6.26-2-686). I've
> started a discussion about my broblem already without any solution.
> It would be interesting for me to know if you partitions are deteced stering
> the driver. Please grep dmsg for your device node and check if they are
> within /proc/partitions too. If not, try to call hdparm -z
> /dev/what-ever-your-device-ist to force the kernel to reload the partition
> table.
> To be able to start hdparm you have to copy the app from somewhere else
> because it's not part of initrd environment. I used an external USB HD witch
> is detected by udev seamless.
>
> My next try will be to go back to an older linux version and udev. Maybe
> udev is also an problem.
>
> My problem is "Resourse or device busy" at this moment within the initrd
> environment afer hparm -z reloaded the partitiontable and udev created the
> device node.
Unfortunately, when this problem occurs, it does not give me any kind of
shell, so it is very hard to get more info (all I can do is hard reboot).
There are several messages regarding ata1, scsi, and sd, that seem to
indicate that the disks are seen at some level, but that is all I can say.
Thanks,
Kelly Clowers
--
To UNSUBSCRIBE, email to debian-us...@lists.debian.org
Andrew Reid
> Is there a way to get a shell inside the initramfs, to
> look around?
You can boot with "break=mount" as a kernel option, it
will drop you into a busybox shell just before mounting
the root directory. The initramfs environment doesn't
inherit environment variables from the boot scripts, but
you can browse the file system and run the scripts and
stuff.
I've run into similar problems when the initramfs
didn't include the right driver module for the disk drive --
the result is that the BIOS can see the drive, and GRUB
can see it, but the kernel can't -- you get symptoms
similar to what you describe, and it's also consistent
with the LiveCD working, as it presumably has a full
kernel.
The solution in that case is to get the right module into
the initramfs, most simply by just listing it by name
in /etc/initramfs-tools/modules and rebuilding the
initramfs.
I haven't mentioned it because it doesn't make sense
in your case -- you're restoring an RMA'd drive on
an old box, where presumably the motherboard SATA
device hasn't changed, so the initramfs you restored
from back-up should work, plus you said you rebuilt the
initramfs several times.
The other way to investigate the initramfs, incidentally,
is to just unpack it somehwere -- it's a cpio archive, you
can google for instructions.
-- A.
--
Andrew Reid / rei...@bellatlantic.net
Kelly Clowers
> On Tuesday 07 July 2009 15:26:54 Kelly Clowers wrote:
>
>> Is there a way to get a shell inside the initramfs, to
>> look around?
>
> You can boot with "break=mount" as a kernel option, it
> will drop you into a busybox shell just before mounting
> the root directory. The initramfs environment doesn't
> inherit environment variables from the boot scripts, but
> you can browse the file system and run the scripts and
> stuff.
Thanks, that might help.
> I've run into similar problems when the initramfs
> didn't include the right driver module for the disk drive --
> the result is that the BIOS can see the drive, and GRUB
> can see it, but the kernel can't -- you get symptoms
> similar to what you describe, and it's also consistent
> with the LiveCD working, as it presumably has a full
> kernel.
> The solution in that case is to get the right module into
> the initramfs, most simply by just listing it by name
> in /etc/initramfs-tools/modules and rebuilding the
> initramfs.
>
> I haven't mentioned it because it doesn't make sense
> in your case -- you're restoring an RMA'd drive on
> an old box, where presumably the motherboard SATA
> device hasn't changed, so the initramfs you restored
> from back-up should work, plus you said you rebuilt the
> initramfs several times.
Yeah, and anyway, it isn't like I am using rare drivers -
it is an Intel MB, and I don't even have LVM or anything.
Anyway, thanks for the advice, I will see what I can do
with it.
> The other way to investigate the initramfs, incidentally,
> is to just unpack it somehwere -- it's a cpio archive, you
> can google for instructions.
Also good to know.
Thanks,
Kelly Clowers
Kelly Clowers
Going to try this now, since break= does not seem to work for
some reason...
Cheers,
Kelly Clowers
>>> is to just unpack it somehwere -- it's a cpio archive, you
>>> can google for instructions.
>>
>> Also good to know.
>
> Going to try this now, since break= does not seem to work for
> some reason...
and when I use
zcat "<archive>" | ( while true; do cpio -i -d -H newc
--no-absolute-filenames --quiet || exit; done )
I get the output:
19036 blocks
cpio: premature end of file
Which is not reassuring. Files and directories are created, but who
knows it all the bits are there?
But further examination shows that there is no ext4 driver
in /lib.
I add ext4 and ahci (even though that .ko was present) to
the module list, and regenerated the initramfs.
And it now boots! But why/how was ext4 not in there?
Oh well, at least it works!
Andrew Reid
That strikes me as an Important Clue, although I'm not
sure what to make of it. Are you *sure* you're running an
initramfs? ("Is it plugged in?") Does your bootloader know
you want an initramfs? How did you get the bootloader on
to the disk? It's not part of the filesystem, so it didn't
come on with the restore-from-backup unless you used "dd"
to image it. For grub, you have to do a "grub-install"
incantation.
Anyways, assuming you *are* using an initramfs, there are
other predefined breakpoints -- they are:
top, modules, premount, mount, bottom, init.
You might try "break=top", that drops into the busybox shell
immediately after the initrd is unpacked, before running any
scripts at all. It's also before modules are loaded, so it
wouldn't be a huge surprise if the disks are missing at that
point, but at least you could check the integrity if your
initramfs environment, and run some stuff manually, maybe.
"modules" breaks before any modules are loaded.
"premount" is the default, and it'll break here if you
just provide a "break" kernel argument, with no value.
Posting to the list, because this info about initramfs
breakpoints should be in more places.
-- A.
--
Andrew Reid / rei...@bellatlantic.net
Kelly Clowers
I may try the break stuff again later, just to see what is
what, for now I am just happy to have a working system.
Thanks,
Kelly Clowers
lee
> On Mon, Jul 13, 2009 at 11:18, Raphael
> Bossek<raphael...@googlemail.com> wrote:
> > My problem is "Resourse or device busy" at this moment within the initrd
> > environment afer hparm -z reloaded the partitiontable and udev created the
> > device node.
That sounds like a hardware problem to me: Cables, or the disk itself.
> There are several messages regarding ata1, scsi, and sd, that seem to
> indicate that the disks are seen at some level, but that is all I can say.
It might help to know what messages you are getting. Maybe they are
trying to tell you what the problem is.
Are you sure that you have all the modules available that are needed
to access/use the disks on your hardware? It clearly says that the
device that should be mounted doesn't exist, and that indicates that a
required module could be unavailable.