During addShard we find any databases that already exist on the shard being added and attempt to add them to the cluster. If this fails for any reason, however, we log a message and continue. This feels like a bad idea as now a) databases that a user might have been expecting to be visible to the cluster are now hidden and b) while I have not verified this, I suspect this could cause problems if a database with the same name is created sometime in the future.