Migrating a linked clone doesn't show as linked clone

[lince]

Hi there,

I think this might be a bug or a feature that is not yet implemented.

I was migrating from a single server to a cluster. Today I migrated a VM that was a linked clone. The process I followed to do the migration was to create the vm and then overwrite the qcow disk file.

The VM works correctly but on the web interface, there is no information about the "backing file". This information is correct in the qcow2 file:

# qemu-img info vm-100-disk-1.qcow2
image: vm-100-disk-1.qcow2
file format: qcow2
virtual size: 10G (10737418240 bytes)
disk size: 482M
cluster_size: 65536
backing file: ../111/base-111-disk-1.qcow2

I guess maybe migrating with a backup this wouldn't happend. Anyway I thought I would mention it, just in case you consider it would be interesting to create a button to refresh the disk information. This way, the webui could read the information from the qcow disk image and update the information shown in the webui.

I found this situation, but I guess any other case where the qcow image gets modified in the command line would affect the information on the webui.

Regards.

 

[Richard]

I think this might be a bug or a feature that is not yet implemented.

I was migrating from a single server to a cluster. Today I migrated a VM that was a linked clone. The process I followed to do the migration was to create the vm and then overwrite the qcow disk file.

That´s not the usual "migrate" action but a backup / restore in a rather tricky way; why not - but we have to take care then by ourselves what happens .....

Or in other words: backing up of linked clones as a clone (which means will be again a linked clone after restore) is simply not implemented.

 

[cot]

Linked-clones are linked to a base VM template. So migrating linked-clones across node needs the base template to be migrated as well. As I found out, Backing up linked clone with vzdump is possible but a restore will turned the linked-clone to a full-clone.

Anyone knows if is is possible to perform an offline migration of linked-clones on a cluster setup? If it is possible, would the migration from one node to another turn the linked clone to a full clone just like the backup do?