Since you mentioned this adding weight I went and adjusted my ksdevice
from eth0 to eth1 and PXE still loads the kernel right but then I can't
get passed the dhcp allocation screen. Its waiting for a link which is
not hot. So I think the kernel and pxe understand that eth0 is eth0 it
just for some reason fires off too early, or needs to understand
failure/latency and retry some set number of times or have fallback
mechanism.
It needs to get a bit smarter, one point of failure cannot be an
acceptable option.
-a
Lambert, Eric wrote:
I've had this same problem for this reason. i.e. PXE booting a dell
1750 over the on-board gig-1 (chosing between gig-1 and gig-2) with an
additional e1000 card in the box will boot PXE over the gig-1 then fail
on the anaconda DHCP with this error. When I modify the ks.cfg to use
eth1 instead of eth0 it works. As Jeremy stated below, I think it's
because the BIOS and linux differ on which interface is eth0 once the OS
starts to load.
-Eric
-----Original Message-----
From: Jeremy Silver [mailto:jeremy@xxxxxxxxxxxxx]
Sent: Monday, June 07, 2004 12:51 PM
To: 'Discussion list about Kickstart'
Subject: RE: Issue with Aquiring DHCP
Do you have multiple ethernet interfaces? Linux and the BIOS do not
always agree on which is the 'first' ethernet port(and therefore eth0)
in the system. On some systems I have to use the second (as labeled on
the server
chassis) ethernet port for kickstart as Linux see it as eth0, or
specify
eth1 as the ksdevice to use the first ethernet port.
--
Andy Ciordia, Network Engineer
Planned Giving Design Center (www.pgdc.com)
10800-D Independence Pointe Pkwy, Matthews, NC 28105
Ph: (704)849-0731 x106 | Fax: (770)456-5239
--
