If you really think your article now has no value whatsoever, then you can withdraw your arXiv submission, following the instructions on arXiv help: To withdraw an article. The article’s arXiv front page will then be replaced by a withdrawal notice; you can include a comment explaining the reason. (This is basically the official preferred alternative to your “upload a dummy pdf” approach.)
However, it’s very rare that an article is really completely valueless! Chances are it has some differences from the overlapping work, in the approach or the viewpoint or the technical details; if nothing else, it gives an alternative exposition of the material. If on reconsideration you think this is the case, then just upload a revised version, prominently explaining the extent to which it’s a rediscovery of earlier work (definitely in the abstract and/or introduction, perhaps also in the arXiv comments field), and if you feel it’s appropriate, also highlighting what is different or novel in your version.
With both these courses of action, the previously submitted version(s) will still be available if anyone really wants to look at them — that’s essentially unavoidable. But under normal circumstances, it’s extremely unlikely anyone would go looking for them, and even more unlikely that they’d hold it against you.