Home / ubuntu / Questions / 397485
Accepted
Martin Drozdik
Asked: 2013-12-30 18:36:27 +0800 CST

What to do when I get an "attempt to read or write outside of disk 'hd0'" error and Boot Repair does not solve the problem?

  • 772

I tried to install Xubuntu 13.10 on an older computer. I noticed some difficulties during installation that may be relevant to my problems. Notably:

  1. The screen resolution was extremely low. The windows were much larger than the screen and I had to move them around to get to the "Next" button.

  2. The mouse cursor was a blur of colours that was flashing all the time.

The installation finished and I tried to reboot the computer. Immediately I got:

error: attempt to read or write outside of disk 'hd0'. 
Entering rescue mode... 
grub rescue>

I made a Boot Repair USB disk, booted it and chose the default repair. The application told me that the problem is repaired and that If this does not solve my problem I should show this link to someone who can help me.

What should I do next? I already tried to install Ubuntu and Xubuntu, both with no success.

EDIT:

I found a workaround.

  1. Insert the Xubuntu live USB
  2. Select "try Xubuntu"
  3. In the settings menu select "install Xubuntu"

This fixed the problem for me. Since it is solved I cannot test the answer.

grub2

9 Answers

  1. Best Answer

    Locate the partition in which linux is present with the help of following technique

    grub rescue > ls
(hd0) (hd0, msdos9)
grub rescue > ls (hd0,msdos9)/
grub rescue > ls (hd0,msdos8)/
grub rescue > ls (hd0,msdos5)/ # suppose this is root and bootloader of linux
grub rescue > ls (hd0,msdos5)/
grub rescue > set root=(hd0,msdos5)
grub rescue > set prefix=(hd0,msdos5)/boot/grub
grub rescue > insmod normal
grub rescue > normal

    Now, system's boot menu appears. Boot into linux.

    sudo update-grub
sudo grub-install  /dev/sda # If the drive is hd0 the equivalent is sda, if it's hd1 then use sdb

    This fixes boot loader.

    • 40

  2. For anyone else that has this issue:

    This was happening to me where a new install or grub repair would work and reboot, but the next time I rebooted, it would get this same "attempt to read error". Most times I would get a kernel panic and the computer would need a hard restart.

    I followed some advice and did a new install using the advanced options (don't use the "erase disk and install ubuntu") using these partition settings:

    1. Create a 1 GB (1024 MB) ext4 partition on the beginning of the disk; mounted in "/boot"
    2. Create your desired install space in ext4 mounted in "/" MINUS your swap area
    3. Use remaining space for swap. (ALL partitions will be primary)
    4. In the boot install dropdown menu, select your "/boot" partition. Not the defaulted drive root!
    • 30

  3. It can be easily solved through the Grub rescue prompt. The first answer isn't quite complete and I got the same error at first. Here is how it works:

    1st we need to find the primary partition. This will be where the essential files needed for linux to boot will reside, so we enter this -

    Grub> ls

    You should now see a list comparable to (hd0) (hd0,msdos5) (hd0,msdos1). Note that you may have different partitions than myself, as well as multiple drives, but the process is the same.

    Now search the partitions to find the primary -

    Grub> ls (hd0)
Grub> ls (hd0,msdos5)
Grub> ls (hd0,msdos1)

    Until grub outputs either a list of files on that partition or shows the filesystem type and date of last modification.

    2nd we need to set up a few things once we've found our primary partition -

    Grub> set prefix=(hd0,msdos1)/boot/grub
Grub> set root=(hd0,msdos1)
Grub> set

    Once the last set command is executed Grub will output a list of different parameters. Now lets check that we've set the correct root and prefix -

    Grub> ls /boot

    Grub should output a list of files contained within /boot.

    3rd we need to set the mount point and load the kernel -

    Grub> linux /vmlinuz root=/dev/sda1
Grub> initrd /initrd.img

    And lastly we boot the system -

    Grub> boot

    If you end up booting into the busybox shell, simply enter fsck /dev/sda1. After that just enter exit and your system will boot normally.

    • 10

  4. I solved this problem by moving my Linux partition close to begining of hard disk. When linux partition was at the end of my 120 Gb HDD (40 Gb from end) I got this problem. No one solution didn't work. After moving linux partition to space from 40Gb to 80Gb problem disappeared.

    • 3

  5. I had this issue when trying to install Ubuntu 18 Server and Xubuntu 18 on a Dell server. Selecting UEFI instead of Legacy in the BIOS menu fixed it.

    • 3

  6. After searching lots of hours I tried this solution and it works well, to reinstall the kernel:

    sudo dpkg --configure -a
sudo apt-get install --reinstall linux-image-$(uname -r)
    • 2

  7. This is random problem.

    In this case better solution is to create /boot folder in its own partition

    Details are here

    • 1

  8. I am achieving this issue when I has used vmdk disk from OVA virtual machine package in virt-manager/QEMU/KVM. The virtual machine was failing with this error message.

    Solution was to convert vmdk into qcow2:

    qemu-img convert -O qcow2 MyAppliance-disk1.vmdk MyAppliance.qcow2

    And to use converted image in virt-manager instead.

    • 0

  9. I used @Andrew answer to recover an existing system.

    In my case I had the following system:

    • /sda1 primary ntfs Windows Boot
    • /sda2 primary ntfs Windows System
    • /sda5 extended ext4 Old Ubuntu Root
    • /sda6 extended swap

    Basically I installed a new ubuntu besides the old one with an independent /boot partition, then remapped the old ubuntu to use the new /boot.

    1. Start Installing new ubuntu from live CD, go to advanced.

    2. Make space for a /boot partition (permanent) and a root partition (temporary) for the new ubuntu.

    I resized an ntfs partition to make a 4gb /boot primary partition and deleted the swap partition to make the new root (/) partition. /boot might not need to be primary, and everything could be done inside the repurposed swap without resizing. This was my partition table after this:

    • /sda1 primary ntfs (Windows Boot)
    • /sda2 primary ntfs (Windows System) (shrinked)
    • /sda4 primary ext4 /boot
    • /sda5 extended ext4 / (Old Ubuntu Root)
    • /sda6 extended ext4 / (New Ubuntu Root)

    1. After installing you might access the old ubuntu at grub menu below the memtest (it's recognized automatically like if it was a windows partition). Boot it.

    2. Edit /etc/fstab to mount the new /boot, ie

    UUID=2d7611ad-0547-4d7c-a595-3f4103ff0b0e /boot ext4 defaults 0 2

    1. Add your old kernel to the new boot partition:

    • Option 1 (tested) Mount the new /boot and reinstall all linux-images

      mount /boot

      apt-get install $(apt list --installed | grep linux-image | cut -d / -f -i)

    • Option 2 (not tested) Copy everything from your old /boot to the new one, except the grub folder

      cp -dR /boot /tmp

      rm -r /tmp/boot/grub

      mount /boot

      cp /tmp/boot/* /boot

    1. Delete the new ubuntu partition and recreate the swap partition. Use gparted for graphic ui. This is my partition table after this
    • /sda1 primary ntfs (Windows Boot)
    • /sda2 primary ntfs (Windows System) (shrinked)
    • /sda4 primary ext4 /boot
    • /sda5 extended ext4 / (Old Ubuntu Root)
    • /sda6 extended swap

    1. Use blkid to find the UUID of your new swap partition, update fstab with the new UUID

    2. Update Grub with update-grub and reboot.

    3. If everything is still working, umount /boot. delete everything inside your old /boot and remount it.

    • 0