On Thu, 2005-05-26 at 10:22 -0400, Alex Kirk wrote:
<SNIP>
> Internet6:
> Destination      Gateway            Flags
> default          ::1                UG
> default          ::1                UG
<SNIP>
> Further attempts at route deletion result in:
> 
> schnarff.com:~$ sudo route delete -inet6 default
> writing to routing socket: No such process
> delete net default: not in table

Try "route delete -inet6 default gw ::1" to remove them.

> This looks thoroughly broken, but as I'm not the IPv6 expert here, I don't 
> know
> how to fix it.

man route :)

> > Try pinging  2001:5c0:8fff:fffe::28f5, 2001:5c0:8fff:fffe::28f4 etc and
> > then try something remote, or just try something remote and see if that
> > works. If you can't ping the ::28f4 then your tunnel is broken, use
> > tcpdump on the IPv4 interface (fxp0 in your case) to see if you get any
> > packets, like proto-41 unreach back from the remote side or from
> > intermediate routers. Or if you get packets back but the kernel filters
> > them out -> firewall issue.
> 
> I can't ping the ::28f4 address. When I run tcpdump (which I have to do on 
> gif0,
> not fxp0, if I want IPv6 traffic), I get:
> 
> schnarff.com:~$ sudo tcpdump -n -i gif0
> tcpdump: WARNING: gif0: no IPv4 address assigned
> tcpdump: listening on gif0
> 10:12:37.890333 2001:5c0:8fff:fffe::28f5 > 2001:5c0:8fff:fffe::28f4: icmp6: 
> echo
> request
> 10:12:38.890316 2001:5c0:8fff:fffe::28f5 > 2001:5c0:8fff:fffe::28f4: icmp6: 
> echo
> request
> 10:12:39.890308 2001:5c0:8fff:fffe::28f5 > 2001:5c0:8fff:fffe::28f4: icmp6: 
> echo
> request
> 10:12:40.890305 2001:5c0:8fff:fffe::28f5 > 2001:5c0:8fff:fffe::28f4: icmp6: 
> echo
> request

As I mentioned, dump fxp0 as now you don't see which source/dest IPv4
you are using and neither are you seeing any ICMP (v4) proto-41
unreaches if the remote side actually doesn't like you.

> Looks like the other side isn't paying any attention to me. Of course, seeing
> this, I noted that ::28f5 appeared to be where I was coming from, so I tried
> setting that as my default route. At that point, I could ping myself (at
> ::28f5), but I couldn't hit, say, 2001:200:0:8002:203:47ff:fea5:3085:

How exactly did your routing table look like after you did exactly what?

You should end up with something like:

default 2001:5c0:8fff:fffe::28f4
2001:5c0:8fff:fffe::28f5 :: dev gif0
2001:5c0:8fff:fffe::28f4 2001:5c0:8fff:fffe::28f5 dev gif0

> schnarff.com:~$ ping6 www.kame.net
> PING6(56=40+8+8 bytes) 2001:5c0:8fff:fffe::28f5 -->
> 2001:200:0:8002:203:47ff:fea5:3085
> ping6: sendmsg: No route to host
> ping6: wrote www.kame.net 16 chars, ret=-1
> ping6: sendmsg: No route to host
> ping6: wrote www.kame.net 16 chars, ret=-1
> 
> > Good thing about IPv6, you can destroy it and IPv4 keeps working.
> > Alternatively when you have IPv4 and IPv6 native, like me, either of the
> > two can die, get firewalled and it will still work ;)
> 
> I'm well aware of this...I just didn't want to start touching "default" 
> routes,
> since a simple syntax error on my part could result in the whacking of my IPv4
> default route.

IPv6 tools don't touch the IPv4 ones, unless they are severely broken.

> Given this, does the need to have some modern documentation on the subject 
> seem
> a bit more clear? ;-)

TSP client should do it already for you.
For the rest: google(openbsd rc.conf ipv6) eg
http://schvin.net/writings/openbsd-ipv6.html
http://www.fbunet.de/ipv6.shtml

The latter being OpenBSD 3.5, just need to search correctly ;)
As KAME (used on *BSD) is the most used IPv6 stack it really works.

Or dump your route table + interfaces again and do the ping tests, on
the IPv4 interface (fxp0) I mentioned.

Greets,
 Jeroen

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to