[nSLUG] Eastlink cable modem access

Jon me at jonwatson.ca
Thu Jan 31 14:02:08 AST 2008


On Jan 31, 2008 1:51 PM, Gerald <linux at zdoit.airpost.net> wrote:

> Jon wrote:
> . . . .
> > Interesting. Granted, I don't spend a lot of time watching DHCP logs,
> > but I would think that the second example is more indicative of a
> > 'normal' DHCP request as it is a broadcast request on port 67. This is
> > designed to hit any DHCP server on the network in the hopes that it will
> > give out an address.
> >
> > Given that a computer without an IP address cannot communication on a
> > TCP/IP network. I can't see how a DHCP request to a specific address
> > (like  in your first snippet) would succeed. Perhaps it's a refresh
> > request in that the interface already has an IP assigned to it and can
> > therefore query a specific DHCP server.
> >
> > Curious...
> >
> > Jon
>
> Jon,
>
> I wasn't looking at the logs until I noticed a connection from my router
> to the 10.6.x.x address on a DHCP port. Then I started looking at the
> logs.
>
> My understanding of the DHCP protocol is it can/must ask first for a
> renewal from the same DHCP server that offered the address to be
> renewed. That leads to requests to a specific address instead of the
> 255.255.255.255 address.
>

Hmm...dunno.

RFC 2131 doesn't state that.

Section 3.1 deals with client broadcast requests and section 3.2 deals with
client specific requests (as in the client is asking for a specific IP
address). Both these sections list step 1 as 'the client broadcasts as
DHCPREQUEST message...' which would seem to preclude the specification of a
DHCP server as a valid step in the DHCP process.

However, that's a big document and I've spent all of 2 minutes with it so
who knows.

I know none of this helps with your problem. Thanks for humouring my idle
curiousity.

Jon
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://nslug.ns.ca/mailman/private/nslug/attachments/20080131/8fe43f25/attachment.html>


More information about the nSLUG mailing list