> > I've read the arguments, and agree with his rationals for not using > CNAME, but I don't believe that his rational is reason enough to > completely avoid using them. The biggest problem is it can get messy if > CNAME are pointing to CNAME pointing CNAME, and so on. >
Actually I'm pretty sure that multiple levels of indirection are forbidden in CNAME's. There are no recursive lookups when resolving CNAMEs, err, rather, it's never guaranteed. Often it's implemented, but not to be depended on. Iterative lookups ONLY are required to be a function of a resolver, so if you use multiple layers of lookups, for some people it just might not work. I haven't read this paper, but maybe the people who came up with DNS thought it was a good idea to avoid too. -Tom -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
