On Fri, 2005-11-04 at 20:16 -0600, Jay Cliburn wrote: > Thanks a bunch for your reply Mike. It'll take me awhile to digest > it, and I'm sure I'll have more questions (like the one I just sent), > but you've sent me some great info. Yeah... I saw your other message. That's not the road I would go down (it's sort of like trying learn the Internet by setting up a 10.* net for yourself first and playing with it internally - yes you can but should you?) but I can (and will) work with it. I'll work up some example configurations for you (and everyone else) that should work for both private space (site locals) and the global unicast. Basically, the only difference will be the "prefix", so I can easily generalize that. I'll try and post a more elaborate follow-up with examples tomorrow. I'll reply to that other message tomorrow. Basic watchword, however... Don't think in terms of IPv4. IPv6 is NOT IPv4 with fat addresses. There are different ways of working with things. You can, but you shouldn't, assign addresses to machines the same way you do (did) in v4 space. Rather, think in terms of a "router" (whether it's actually routing or not) which is advertising a prefix and routes and is authoritative in declaring the network address prefix to the end hosts. The other systems simply fall in line behind this and configure their addresses based on that advertised prefix. This is what we call "stateless autoconfiguration". Different paradigm. Your configuration efforts center around one (or few) machine(s) (the routers) and there is actually very little that needs to be done on the end nodes above and beyond simply enabling it. Once you realize that, IPv6 actually becomes easier than IPv4. Although... If you are not working with a solid DNS, this CAN get more complicated because the autoconfigured addresses are complicated (they incorporate the MAC address). You're going to have to deal with ugly addresses and bury them in DNS or in /etc/host files and work on a name level or you will be struggling. When ever you have to resort to IPv6 address literals, you are, most likely, doing something wrong. > Thanks again. I really appreciate it. FWIW, I've got ping6 working > using link-local addresses. Now I'm trying to set up site-local > addresses on each host, but I'm running into a little trouble, hence > the request for ifconfig and ifcfg-eth0 info. If you are statically assigning addresses to each machine, that's the first thing we have to correct. You're doing way too much work (that you'll just have to undo anyways). Don't go there. There are a couple of easy setups that will take care of this without going to that much work. And they will work even better once you're on the v6 net with a global unicast prefix. > Jay > > On 11/4/05, Michael H. Warfield <mhw@xxxxxxxxxxxx> wrote: > > [excellent info deleted for brevity] > Mike -- Michael H. Warfield | (770) 985-6132 | mhw@xxxxxxxxxxxx /\/\|=mhw=|\/\/ | (678) 463-0932 | http://www.wittsend.com/mhw/ NIC whois: MHW9 | An optimist believes we live in the best of all PGP Key: 0xDF1DD471 | possible worlds. A pessimist is sure of it!
Attachment:
signature.asc
Description: This is a digitally signed message part
