Goal

• As A KubeVirt user, I want to be able to perform volume migration from a block to filesystem/block and keeping the image sparse.

Currently, when a block volume is migrated, it is fully copied, see upstream GH issue [1].

Acceptance criteria

• Be able to detected zeros on a block device during the storage migration
• Keep the destination image sparse 

[1]https://github.com/kubevirt/kubevirt/issues/12876

Additional information

source XML:

    <disk type='block' device='disk' model='virtio-non-transitional'>
      <driver name='qemu' type='raw' cache='none' error_policy='stop' io='native' discard='unmap'/>
      <source dev='/dev/datavolumedisk1' index='1'/>
      <backingStore/>
      <target dev='vda' bus='virtio'/>
      <alias name='ua-datavolumedisk1'/>
      <address type='pci' domain='0x0000' bus='0x07' slot='0x00' function='0x0'/>
    </disk>

destination XML:

    <disk type='block' device='disk' model='virtio-non-transitional'>
      <driver name='qemu' type='raw' cache='none' error_policy='stop' io='native' discard='unmap'/>
      <source dev='/dev/datavolumedisk1' index='1'>
        <slices>
          <slice type='storage' offset='0' size='5368709120'/>
        </slices>
      </source>
      <backingStore/>
      <target dev='vda' bus='virtio'/>
      <alias name='ua-datavolumedisk1'/>
      <address type='pci' domain='0x0000' bus='0x07' slot='0x00' function='0x0'/>
    </disk>

ceph pool information:

$ rbd du --pool=replicapool
NAME                                          PROVISIONED  USED   
csi-vol-739d5db3-86f5-11ef-b8b5-f2bd1cccbeca        5 GiB   64 MiB
csi-vol-bf1b7510-86f5-11ef-b8b5-f2bd1cccbeca        5 GiB    5 GiB
<TOTAL>                                            10 GiB  5.1 GiB

The 'csi-vol-739d5db3-86f5-11ef-b8b5-f2bd1cccbeca' is the source volume while the 'csi-vol-bf1b7510-86f5-11ef-b8b5-f2bd1cccbec' is the destination. The bug result consists that the destination volume is fully allocated, while we expect that it should have the same used storage as the source.