Grub error: no such disk

Simone-

Was this in a VM, or on real hardware?

Also, can you tell us about your disks? What size? SATA, SCSI, or IDE?

Some more info: the system is a DELL SC1425 with 2 scsi disks; i've set up 2 md devices: md0 (~1GB, used for swap) and md1 (~70GB, formatted as EXT4, mount as /).
After the reboot at the end of the installation, the system halts.
Grub says "no such disk" and show me the "grub rescue" prompt.

If I type "ls <return>", grub show me:
(md0) (hd0) (hd0,2) (hd0,1) (hd1) (hd1,2) (hd1,1)

Note that (md1) is missing.

I cannot perform any other command short of "ls"; "boot", "root", "kernel" are not working from this mode.

In data martedì 27 aprile 2010 hai scritto:
> Simone-
>
> Was this in a VM, or on real hardware?
>
> Also, can you tell us about your disks? What size? SATA, SCSI, or IDE?
>
> ** Changed in: grub (Ubuntu)
> Importance: Undecided => High
>
> ** Package changed: grub (Ubuntu) => grub2 (Ubuntu)
>
> ** Changed in: grub2 (Ubuntu)
> Importance: High => Undecided
>
> ** Changed in: grub2 (Ubuntu)
> Importance: Undecided => High

I've added some more info on the launchpad bug; anyway the system is a real
(no VM) dell sc1425, with 2 scsi disks. 73Gb each, if I recall correctly.

I've set up 2 md devices: md0 (~1GB, used for swap) and md1 (~70GB, formatted
as EXT4, mount as /).

After the reboot at the end of the installation, the system halts.
Grub says "no such disk" and show me the "grub rescue" prompt.

If I type "ls <return>", grub show me:
 (md0) (hd0) (hd0,2) (hd0,1) (hd1) (hd1,2) (hd1,1)

Note that (md1) is missing.

I cannot perform any other command short of "ls"; "boot", "root", "kernel" are
not working from this mode.

I've rebooted with a system rescue cd which recognised both the md devices.
Mounted /dev/md1 without problem, but cannot use the installed system (I'd
like to chroot into the /dev/md1 partition to try to reinstall grub), cause
I've installed an x86_64 version while the system rescue cd is only 32 bits.

--
Simone Lazzaris
   INTERACTIVE NETWORK SRL
   Via Roggia Vignola 9, 24047 Treviglio (BG)
   tel : +39 0363.302820
   fax : +39 0363.304352
   web : http://www.interactive.eu
   email : <email address hidden>

Can you boot in rescue mode and attach /var/log/installer/syslog and /var/log/installer/partman from the installed system to this bug, please?

(The ISO image you installed from should have a suitable rescue mode that matches the architecture of your installed system; no need to resort to a separate rescue CD!)

FYI, based on my testing results in http://iso.qa.ubuntu.com/qatracker/result/4133/286, this seems to be hardware specific (ie, no unexpected errors with KVM/virtio disks with various install configurations).

I was not able to reproduce this either, in a VM.

I'm attaching a screen shot of the partitioning, which I tried to mimic as closely as possible to yours (md0 = 1G swap, md1 = ~70G ext4 /). I used scsi emulation instead of virtio.

Install completed and I rebooted successfully.

However, I'm seeing weirdness of RAID1 on Lucid on weird hardware too. I can't confirm your bug, but I have opened another one, where my RAID on real hardware fails to mount /dev/md0 on /root in initramfs.

Colin,
here are the files you have requested:this is the syslog

and this is the partman

Dustin,
the partitioning is correct, exept that I didn't used extended partitions; in other words, md0 is composed of hda1 and hdb2, md1 is composed of hda2 and hdb2, not hda5 and hdb5. Oh, and the md devices were already present on the disks: I've recycled an redhat installation, reformatting md1 but without rebuilding the raid from scratch.

I'm thinking to retry the installation on the system, using a different partition scheme. I've noted that md0 is recognised by grub, so I'll try to use 3 md devices: the first will be used to host the /boot filesystem, the second will be the swap space, the third will host /.

I'll post the results here.

What happened Simon ??

I have the same problem ..
but am beginner linux user..n i don't understand all this things you are talking about !!

I've wiped the disk (with dd if=/dev/zero of=/dev/sda and dd if=/dev/zero of=/dev/sdb) and then retried the installation.
This time, I've builded 3 md devices: md0 at the beginning of the disks, to hold the /boot partition, md1 to hold the swap and md2 for /. I've also formatted /boot as ext3 (not ext4) and / as ext4.

This is the partition scheme for sda (sdb is identical)
root@gonzalez:~# LANG=EN_en fdisk -l /dev/sda

Disk /dev/sda: 73.5 GB, 73543163904 bytes
255 heads, 63 sectors/track, 8941 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x0004a136

   Device Boot Start End Blocks Id System
/dev/sda1 * 1 19 145408 fd Linux raid autodetect
Partition 1 does not end on cylinder boundary.
/dev/sda2 19 140 976896 fd Linux raid autodetect
Partition 2 does not end on cylinder boundary.
/dev/sda3 140 8942 70695936 fd Linux raid autodetect

With this scheme, build from scratch within the installation, all went smoothly and the system is now in production.
I think the problem was related to
1) /boot formatted as ext4 and/or
2) the partition scheme was build in a previous redhat installation. I've reformatted the filesystems but not rebuild the md array in the first installation.

[Expired for grub2 (Ubuntu) because there has been no activity for 60 days.]