Comment 0 for bug 1372472

The vm system disk data was overided by glance image data when evacuating a vm on shared storage.

The nova/compute/manager.py file in icehouse version:

    @object_compat
    @messaging.expected_exceptions(exception.PreserveEphemeralNotSupported)
    @wrap_exception()
    @reverts_task_state
    @wrap_instance_event
    @wrap_instance_fault
    def rebuild_instance(self, context, instance, orig_image_ref, image_ref,
                         injected_files, new_pass, orig_sys_metadata,
                         bdms, recreate, on_shared_storage,
                         preserve_ephemeral=False):
     **********
            kwargs = dict(
                context=context,
                instance=instance,
                image_meta=image_meta,
                injected_files=files,
                admin_password=new_pass,
                bdms=bdms,
                detach_block_devices=detach_block_devices,
                attach_block_devices=self._prep_block_device,
                block_device_info=block_device_info,
                network_info=network_info,
                preserve_ephemeral=preserve_ephemeral)

the kwargs lost parameter recreate=recreate, and it result in the data always destroyed when a vm evacuated.