Edit

Share via

Resolve "Failed to start Switch Root" for an Azure Linux VM

Applies to: ✔️ Linux VMs

This article shows how to resolve the error "Failed to start Switch Root" for an Azure Linux Virtual Machine (VM). This issue can occur when you update to the following GRUB package versions on Oracle 8 SP2:

  • grub2-common-2.02-78.0.3.el8_1.1.noarch
  • grub2-pc-modules-2.02-78.0.3.el8_1.1.noarch
  • grub2-efi-x64-2.02-78.0.3.el8_1.1.x86_64
  • grub2-tools-extra-2.02-78.0.3.el8_1.1.x86_64
  • grub2-tools-2.02-78.0.3.el8_1.1.x86_64
  • grub2-pc-2.02-78.0.3.el8_1.1.x86_64
  • grub2-tools-minimal-2.02-78.0.3.el8_1.1.x86_64

If your Azure issue is not addressed in this article, visit the Azure forums on MSDN and Stack Overflow. Post your issue in either these forums or to @AzureSupport on Twitter.

To submit an Azure support request on the Azure support page, select Get support.

Manually fix the issue in the Guest OS

If you have access to the Azure Serial Console proceed with the steps in the following Azure Serial Console method section.

If you're unable to use the Azure Serial Console section, proceed to the Offline method.

Azure Serial Console method

  1. Reboot the VM using the Azure Serial Console, and hold the Escape key to access the GRUB menu.

  2. Select the entry for the rescue kernel.

  3. Copy the kernelopts value from the grubenv file. The path of the grubenv file in Linux can vary depending on the distribution and configuration of the system. It is commonly located at /boot/efi/EFI/redhat/grubenv or /boot/grub2/gubenv.

    Example:

    cat /boot/grub2/gubenv

  4. Edit the desired boot entry from /boot/loader/entries:

    Example:

    vi /boot/loader/entries/a358b364a6d3492898bedc8d1dea3e92-4.18.0-147.8.1.el8_1.x86_64.conf

    Manually add the kernelopts value that you copied and earlyprintk=ttyS0 to the boot entry file.

  5. Reboot the VM. If more than one kernel is installed, you might need to select the modified entry from the GRUB menu.

Offline method

If you're unable to access the VM using the Azure Serial Console, then the repair must be done in offline mode, as the VM isn't booting.

  1. Use steps 1-3 of the VM Repair Commands to prepare a Repair VM.

  2. Using SSH, connect to the Repair VM.

  3. Once connected to the repair VM, use lsblk to find your boot and efi partitions:

    Screenshot shows the boot and efi partitions.

  4. Create a temporary mount point. For example, use /repair.

    mkdir /repair

  5. Mount the boot partition on the temporary mount point. For example, use /dev/sdc1.

    mount /dev/sdc1 /repair/

  6. Mount the efi partition under efi on the temporary mount point. For example, use /dev/sdc15.

    mount /dev/sdc15 /repair/efi/

  7. Copy the kernelopts value from the grubenv file. The path of the grubenv file in Linux can vary depending on the distribution and configuration of the system. It is commonly located at /boot/efi/EFI/redhat/grubenv or /boot/grub2/gubenv.

    Example:

    cat /boot/efi/EFI/redhat/grubenv

  8. Edit the desired boot entry from /boot/loader/entries:

    Example:

    vi /boot/loader/entries/a358b364a6d3492898bedc8d1dea3e92-4.18.0-147.8.1.el8_1.x86_64.conf

    Manually add the kernelopts value that you copied and earlyprintk=ttyS0 to the boot entry file.

  9. Unmount the efi partition. For example, /repair/efi/.

    umount /repair/efi

  10. Unmount the boot partition. For example, /repair/.

    umount /repair

  11. Use step 5 of VM Repair Commands to rebuild the VM.

  12. Start the VM. If more than one kernel is installed, you might need to select the modified entry from the GRUB menu.

Contact us for help

If you have questions or need help, create a support request, or ask Azure community support. You can also submit product feedback to Azure feedback community.