Modifying the boot configuration for Red Hat OpenShift Virtualization

Modify the boot configuration of the migrated disk to make it compatible with Red Hat OpenShift Virtualization.

1. Attach the disk to the management VM by running the following command:

   virtctl addvolume management_vm_name \
     --volume-name=pvc_name \
     --bus=virtio \
     --persist \
     --namespace=namespace

2. Connect to the management VM.

   Use SSH or the OpenShift Virtualization console to access the management VM.

3. Switch to the root user by running the following command:

   sudo su -

4. Identify the attached disk device by running the following command:

   lsblk

   Example output:

   NAME   MAJ:MIN RM  SIZE RO TYPE MOUNTPOINTS
   vdc    252:32   0 58.9G  0 disk
   ├─vdc1 252:33   0  4.7G  0 part
   └─vdc2 252:34   0 41.9G  0 part

   In this example, the attached disk is /dev/vdc with two partitions: /dev/vdc1 (boot) and /dev/vdc2 (root).

5. Create a mount point for the migration by running the following command:

   mkdir /mnt/zvm_migration

6. Mount the root file system by running the following command:

   mount root_partition /mnt/zvm_migration

   Example:

   mount /dev/vdc2 /mnt/zvm_migration

7. Bind the mount system directories by running the following commands:

   mount --bind /dev /mnt/zvm_migration/dev
   mount --bind /proc /mnt/zvm_migration/proc
   mount --bind /sys /mnt/zvm_migration/sys

8. Mount the boot file system by running the following command:

   mount boot_partition /mnt/zvm_migration/boot

   Example:

   mount /dev/vdc1 /mnt/zvm_migration/boot

9. Enter the chroot environment by running the following command:

   chroot /mnt/zvm_migration

10. Remove z/VM-specific boot parameters by running the following command:

    sed -i 's/ rd\.zfcp=[^ ]*//g; s/ cio_ignore=[^ ]*//g; s/ rd\.znet=[^ ]*//g' /boot/loader/entries/*.conf

    This command removes the following z/VM-specific parameters:

    • rd.zfcp: FCP device configuration
    • cio_ignore: Channel I/O ignore list
    • rd.znet: Network device configuration

11. Get the kernel version by running the following command:

    KERNEL_VERSION=$(uname -r)

12. Regenerate the initramfs by running the following command:

    dracut -f --no-hostonly --kver "${KERNEL_VERSION}"

    The --no-hostonly option ensures that the initramfs includes all necessary drivers for the new environment.

13. Write the bootloader configuration by running the following command:

    zipl -V

    Expected output example:

    Looking for components in '/lib/s390-tools'
    Using config file '/etc/zipl.conf'
    Using BLS config file '/boot/loader/entries/57ed29736ff244a59b5b1b8041d98aa0-5.14.0-611.5.1.el9_7.s390x.conf'
    Using BLS config file '/boot/loader/entries/57ed29736ff244a59b5b1b8041d98aa0-0-rescue.conf'
    Secure boot support: no
    Target device information
      Device..........................: fc:20
      Partition.......................: fc:21
      Device name.....................: vdc
      Device driver name..............: virtblk
      Type............................: disk partition
      Disk layout.....................: SCSI disk layout
    Building bootmap in '/boot'
    Building menu 'zipl-automatic-menu'
    Adding #1: IPL section '(5.14.0-611.5.1.el9_7.s390x)' (default)
      kernel parmline...: 'root=UUID=2f168e2c-8b5c-4f7a-878e-959993079bae crashkernel=1G-2G:192M,2G-64G:256M,64G-:512M'
    Syncing disks...
    Done.

    Verify that the output shows the device driver as virtblk and the disk layout as SCSI disk layout, confirming KVM compatibility.

14. Exit the chroot environment by running the following command:

    exit

15. Unmount all file systems by running the following commands:

    ROOT_MOUNT=/mnt/zvm_migration
    
    umount "${ROOT_MOUNT}/boot"
    umount "${ROOT_MOUNT}/sys"
    umount "${ROOT_MOUNT}/proc"
    umount "${ROOT_MOUNT}/dev"
    umount "${ROOT_MOUNT}"

16. Detach the volume from the management VM by running the following command:

    virtctl removevolume management_vm_name --volume-name=pvc_name

17. Label the PVC as a bootable volume by running the following command:

    oc label pvc pvc_name -n namespace \
      template.kubevirt.io/architecture=s390x \
      instancetype.kubevirt.io/default-instancetype=u1.small \
      instancetype.kubevirt.io/default-instancetype-kind=VirtualMachineClusterInstancetype \
      instancetype.kubevirt.io/default-preference=rhel.9.s390x \
      instancetype.kubevirt.io/default-preference-kind=VirtualMachineClusterPreference \
      --overwrite

    Note: Add labels to identify the PVC as a bootable volume with appropriate architecture and preferences.
18. Create a DataSource for the bootable volume.
    1. Create a YAML file for the DataSource resource:

       apiVersion: cdi.kubevirt.io/v1beta1
       kind: DataSource
       metadata:
         name: datasource_name
         namespace: namespace
       spec:
         source:
           pvc:
             name: pvc_name
             namespace: namespace

    2. Apply the YAML file by running the following command:

       oc apply -f datasource-yaml-file