It might happen that something else is newer then cleanup delivery (eg. because
of user was able to ask to stop/start just after cleanup).

So clone currently waiting cleanup delivery with date newer then previous one
and replace it. Thanks to this instance destruction will be possible.