ISC-DHCP6 does not send replies

[Matthew Seaman]

On 29/08/2010 15:28:30, Indexer wrote:

>> Connecting to [ff02::1:2]:547 (link-scoped
>> > All_DHCP_Relay_Agents_and_Servers) or [ff05::1:3]:547 (site-scoped
>> > All_DHCP_Servers) should get some sort of answer.

> I can ping6 to ff02::1:2 successfully.

Can you tell if it's your DHCP6 server responding?

>> > Check the routing table on server and client -- on a FreeBSD box, I get:
>> >
>> > % netstat -r | grep ff02
>> > ff02::%re0         fe80::e2cb:4eff:fe U           re0
>> > ff02::%fwe0        fe80::1e:8cff:fec2 U          fwe0
>> > ff02::%fwip0       fe80::21e:8c00:c2: U         fwip0
>> > ff02::%lo0         localhost          U           lo0
>> > ff02::%gif0        fe80::e2cb:4eff:fe U          gif0
> 
> Here is my routing table on my gateway system, using the same command as yours.
> 
> ff02::/16                         ::1                           UGRS        lo0
> ff02::%em0/32                     fe80::216:e6ff:fe7f:972e%em0  U           em0
> ff02::%lo0/32                     ::1                           U           lo0
> ff02::%tun0/32                    fe80::216:e6ff:fe7f:972e%tun0 UGS        tun0
> ff02::%tun2/32                    fe80::216:e6ff:fe7f:972e%tun2 U          tun2
> ff02::%tun3/32                    fe80::216:e6ff:fe7f:972e%tun3 U          tun3
> ff02::%tun1/32                    fe80::216:e6ff:fe7f:972e%tun1 U          tun1

> That ff02::/16 does not look quite right .....

It's different, yes.  That could be due to running DHCP6 -- after all,
the daemon has to have some way of receiving all the DHCP traffic to the
various site- and link- local addresses.  You can test that by turning
off dhcpd and checking the routing table with it not running.  If the
route doesn't disappear, try disabling dhcpd in /etc/rc.conf, rebooting
and then see if that route is still present.  Either way, re-enable
dhcpd in rc.conf and re-start the daemon: if the route appears then it's
required by dhcpd and everything looks to be in order.

> In fact, could that be the issue? I have dhcp6c running from my
> pppoe session (tun0), and it assigns the prefix to em0. I also am trying to
> use em0 as the DHCP6 server. This shouldn't be breaking it, but it
> *could* be?

Ah.  Yes, this might cause you problems.  Possibly.  If Internode DHCP6
has been configured as authoritative for your address range and if the
query packets from your client can reach Internode's DHCP6 server then
you probably will have trouble.  I shouldn't think its likely though --
your client's DHCP6 initial queries will be to find a server on the same
network segment, and to reach the Internode servers it would have to hop
through your gateway machine, which is your DHCP6 server anyhow.

If your network prefix is dynamically assigned, then I don't think there
is a way to have a DHCP6 server be a DHCP6 client as well, and pass on
the prefixes it has obtained dynamically.  BICBW.  If your ISPs policy
is actually to assign you a particular prefix permanently rather than
give you one out of some dynamically assigned pool, then it's worth a
try using a static configuration on your gateway machine -- I believe
you said this was a test setup to see if it could be rolled out on a
customer network?  Should be fine to try static configuration like that
for a limited time even supposing it's all dynamically assigned.

	Cheers,

	Matthew

-- 
Dr Matthew J Seaman MA, D.Phil.                   7 Priory Courtyard
                                                  Flat 3
PGP: http://www.infracaninophile.co.uk/pgpkey     Ramsgate
JID: matthew at infracaninophile.co.uk               Kent, CT11 9PW

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 267 bytes
Desc: OpenPGP digital signature
Url : http://lists.freebsd.org/pipermail/freebsd-questions/attachments/20100829/d486a43e/signature.pgp