IPv6 and packet sending
hi, in my proxy.conf I have a FQDN for a proxy destination. that FQDN has a AAAA record (and all other operations to it from the server us IPv6 for transit). however, FreeRADIUS doesnt want to talk to that remote proxy via IPv6 anyone else successfully SENDING proxied packets from FreeRADIUS using IPv6? have you had to change/bork something to make it use IPv6 for the transit instead of IPv4? the FQDN also has an A record...but the RFC states that if theres a AAAA for the target then it should be used in preference to the A record..... alan
Alan Buxey wrote:
in my proxy.conf I have a FQDN for a proxy destination.
That's generally a bad idea, but OK...
that FQDN has a AAAA record (and all other operations to it from the server us IPv6 for transit). however, FreeRADIUS doesnt want to talk to that remote proxy via IPv6
What does that mean? It *prefers* IPv4. But if a FQDN has *only* an IPv6 record, it will use that.
anyone else successfully SENDING proxied packets from FreeRADIUS using IPv6? have you had to change/bork something to make it use IPv6 for the transit instead of IPv4?
the FQDN also has an A record...but the RFC states that if theres a AAAA for the target then it should be used in preference to the A record.....
Which RFC? Not any of the RADIUS ones, IIRC. Alan DeKok.
Hi,
in my proxy.conf I have a FQDN for a proxy destination.
That's generally a bad idea, but OK...
ah really? okay - I can dig that (no pun intended honest!) - it best practice is to use the IPv4/IPv6 address instead then I can go that way.
What does that mean? It *prefers* IPv4. But if a FQDN has *only* an IPv6 record, it will use that.
I'd already tested that theorem and found that yes, if it only had a AAAA then it'd get used.
Which RFC? Not any of the RADIUS ones, IIRC.
its a more general protocol thing and not a specific one: RFC 3484 - Default Address Selection for Internet Protocol version 6 (IPv6) in particular 10.3 - "This means that applications will use IPv6 in preference to IPv4 when the two are equally suitable" this statement has been, over the past few years, found to be generally true for other services and daemons... which was why what I found earlier this week with FR was a little peculiar. alan
Alan Buxey wrote:
Hi, in particular 10.3 - "This means that applications will use IPv6 in preference to IPv4 when the two are equally suitable"
this statement has been, over the past few years, found to be generally true for other services and daemons... which was why what I found earlier this week with FR was a little peculiar.
OK... that's easy enough to fix. Alan DeKok.
participants (2)
-
Alan Buxey -
Alan DeKok