That being said, the ugly truth of the matter is that there is no agreed-on standard for "how much is enough." People can get hired at Research I campuses with no publications whatsoever, but get tossed out of a "lesser" school's search pool for precisely that reason. A campus with a 5-5 teaching load may look askance at someone who has already knocked out four articles, while one with a 3-3 may be pleased. Publishing a seminar paper that's out of your field may earn your brownie points over here, but eliminate you over there. Nevertheless, most search committees don't expect or even want a graduate student who has already been responsible for the death of several trees, especially if said graduate student has published in journals with poor or non-existent reputations.
Now, "should graduate students be publishing?" is an entirely different question altogether. In an ideal world, my answer, in fact, would be "not until they're near the end of their dissertation research." (Before you ask: no, I didn't publish anything as a graduate student, although I did attend a couple of conferences.) This isn't an ideal world, unfortunately, and efforts to alter the current state of affairs don't seem to be catching fire.