Were you able to resolve this problem without deleting and re-creating your target dataset? I just posted a similar issue and a possible workaround using zfs from the command line. Waiting to hear back before trying to use zfs directly though as I’d really like to avoid messing up my target dataset.