The patch is worth committing because localhost.localdomain is a valid variant. However, I can't find anywhere why some machines are configured like this. Does anyone know? A question was raised, however, as to why we've not run into this issue before, since we rely on identifying localhost in other parts of the code. If we haven't failed before on this type of check, why now? Amareshwari, do you happen to know how your machine got configured like this?