For the past eight weeks or so, Google Groups occasionally failed to show sci.math.symbolic posts created by means of the same interface, although these posts were immediately passed on to usenet (at least some of them also showed up in Google Groups a few days later). Users were thus induced to issue multiple posts, as has happened to Vladimir, Albert, and Dave.
By contrast, posts made using the Math Forum interface run by Drexel University always seem to appear on their list (along with posts generated elsewhere), but are never passed on to usenet. This has happened to at least two recent messages by Fermat author Robert Lewis, which never reached the Italian aioe.org server I am relying on. I have verified the malfunction by means of a dummy post to the "Rubi 3 Usage?" thread today.