I mirrored a large repository of files (CTAN) and was able to get them archived into a gzipped tarball. This file happens to be several gigabytes in size.
I moved that file to a directory serving web pages, and tried to download it over http, but permission was denied. I checked the permissions of the file, and they are the same as any other file that can successfully be downloaded from that directory.
Is there some specific filter against tarball downloads, or downloads of a certain size?