Question on address configuration

Nick Sayer nsayer@quack.kfu.com
Wed, 16 Jan 2002 07:13:50 -0800


Robert Elz wrote:

>
>Certainly DNS servers, DHCP servers, routers, and even SMTP servers (thanks
>all the same, but I don't want a 2 day mail dead spot while my old DNS
>records TTL times out...) are the most likely candidates for configured
>addresses, but any host can have it.   Implementations that I'm aware of
>all allow this already - all they don't do is the compromise where the
>prefix comes from RA messages (just like fully autoconfigured addresses),
>and the token comes from configuration.   This really should be added.
>
>kre
>
Everything you said is true, but I would add the small postscript that 
in 99% of the cases where you would find it truly desirable to set up a 
well known address, you would be better off making that address 
site-local rather than based on the (routable) prefix. Your arguments 
about an SMTP server's DNS TTL can be made about any DNS records -- if 
the TTLs are causing you heartburn you should lower them. People 
typically want to configure manual addresses so that they don't have to 
keep rewriting resolv.conf files (or equiv). The best way to handle that 
is to look at the latest DNS autolocation draft (I forget the name 
offhand), which suggests fec0:0:0:ffff::[1,2,3]. Doing this you won't 
have to *ever* change references to the server even if your prefix changes.

You are right, however, in that since the spec allows even pointless 
silly behavior implementations should be prepared for it. Perhaps those 
implementations that use rtsol (kame) can modify rtsol to provide an 
additional argument to allow you to specify the suffix.