Re: New dhclient broke multiple domains in domain-name

On Thu, Jun 16, 2005 at 01:12:51PM -0400, John Baldwin wrote:
> [ Apologies if this has already been brought up, I'm still 1800 messages 
> behind on current_at_. ]
> 
> A feature of both the old and new dhclient(8) is that it would take whatever 
> was in the domain-name option returned by the DHCP server and stick it in the 
> 'search' line in /etc/resolv.conf.  Thus, if you wanted to have DNS search 
> multiple domains, you could just pass a space separated list of domains to 
> search in domain-name and it would just work.  I've made use of this 
> "feature" in several different environments in the past including my current 
> test lab.  It's even used in the example dhclient.conf in dhclient.conf(5):
> 
>        interface "ep0" {
>            send host-name "andare.fugue.com";
>            send dhcp-client-identifier 1:0:a0:24:ab:fb:9c;
>            send dhcp-lease-time 3600;
>            supersede domain-name "fugue.com rc.vix.com home.vix.com";
>            prepend domain-name-servers 127.0.0.1;
>            ...
>        }
> 
> The new dhclient is barfing on my domain-name setting now because it doesn't 
> look like a domain name:
> 
> Setting hostname: deimos.baldwin.cx.
> fxp0: link state changed to UP
> DHCPDISCOVER on fxp0 to 255.255.255.255 port 67 interval 8
> DHCPOFFER from 192.168.0.1
> Bogus Host Name option 15: baldwin.cx freebsd.org atl.weather.com (baldwin.cx 
> freebsd.org atl.weather.com)
> Invalid lease option - ignoring offer
> packet_to_lease failed.
> DHCPDISCOVER on fxp0 to 255.255.255.255 port 67 interval 10
> DHCPOFFER from 192.168.0.1
> Bogus Host Name option 15: baldwin.cx freebsd.org atl.weather.com (baldwin.cx 
> freebsd.org atl.weather.com)
> Invalid lease option - ignoring offer
> packet_to_lease failed.
> ...
> 
> I'd very much like the old behavior restored if possible, or an alternative 
> way to achieve the same result (multiple domains in the 'search' part 
> of /etc/resolv.conf).  Note that the old domain-name trick has worked all the 
> way back to at least 4.1 and maybe even back in the 3.x days IIRC.

I've disabled validation of this field which should work around this
issue for now.

-- Brooks

-- 
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529  9BF0 5D8E 8BE9 F238 1AD4