Comment 4 for bug 397188

This is an original design feature of the Librarian. If there are any unexpired URLs referencing the same file on disk, all URLs continue to work. Also, even if all the URLs do expire and the physical file removed from disk, reuploading the file again will allow all the expired URLs to start working again.

If we need to remove security sensitive information, we need to actually remove it rather than just flagging data as expired. This currently involves removing the LibraryFileAlias record if we need to stop a single URL referencing data to stop working, or the LibraryFileContent and all LibraryFileAliases (and deal with all the fallout) if we need to remove all traces from our system for legal reasons.

We certainly don't want removal of the LibraryFileContent being the default.