Re: Grrrr ... dhcpd6

[Pavel Simerda <psimerda redhat com>]

> From: "Gene Czarcinski" <gene czarc net>
> On 09/28/2012 08:15 AM, Pavel Simerda wrote:
> > You*may*  be right. But, unfortunately, I have been playing with
> > DHCPv6
> > implementations and there wasn't one that I would actually like.
> > Except
> > ISC DHCP which works for me but needs improvement.
> Give dnsmasq a look.

AFAIK dnsmasq is out of question here as it only provides the server side.

> It seems that not only NetworkManager using it as a caching name server

And we might be abandoning it, switching to unbound, but that's a completely
different topic. For NetworkManager, the choice of caching/validating
nameserver is orthogonal to the choice of DHCP client anyway.

> but libvirt is using if not dependent on it. Granted, it is nothing I would
> like trying to scale to "industrial size" networks, but it sure configures
> and works well.

We're using it as a local caching nameserver because it's simple to set up. We're
using it as a caching nameserver and DNS server for IPv4 method=shared for the
same reason.

> in another message, there is sudden activity on the dnsmasq discussion
> mailinglist to get its RA support to work better ... maybe even suplant
> the need for radvd

We're not using radvd with NetworkManager at all as we currently don't support
IPv6 connection sharing and we actually don't know yet how to do that. Ask Dan
Winship about this.

> to configure routes when dhcp supplies the address and dynamic dns.

This choice would be made when (if) we add support for IPv6 connection sharing. But
unfortunately for us and for this feature to be added, this won't work as easily
as it does for IPv4.

> You might want to consider adding it to the suite of tests you use. For my testing,
> I am using both named/dhcpd/dhcpd6 and dnsmasq (both
> requiring radvd).

Actually, the reationale for creating my tests was to *have a working IPv6 network
with all sorts of dynamic configurations of host addresses*. To fullfill this I don't
need anything of this sort unless I want to test your dynamic DNS stuff. And as you
are testing it already, I'll kindly ask you to publish your tests.

If you would like to participate on a broader range of interoperability tests, I'll be
only happy. With Dan Williams, we are still looking for a reasonable way to do integration
testing using virtualization. And I'm also working on a way to do complex non-integration
tests for NetworkManager that would not require virtualization nor root account. These
could complement each other.

Cheers,

Pavel

> 
> Gene
>