Network cards failing to find network

[MrTom]

i've recently moved my hdd and network cards into a new machine but the networking has failed.
the one network card goes to the external network and the other is internal.
when i boot the external card is supposed to get an IP address thru DHCP but this fails. the second card is just supposed to come up on a set IP address but also fails.
if i do an ifdown-ifup on the internal card then that network starts working but the external card never works. it uses PUMP but i don't really understand the difference between this and using dhclient (which also doesn't work). both network cards are the same make and use the tulip driver and both worked in the old machine with the same hdd - so nothing has changed but the motherboard (from an old p200 to a p400 based board).
i'm pretty sure that the external network is acknoledging the DHCP request but the kernel never seems to pick it up.

thanks.

 

[ifincham]

Hi,

Well, assuming you properly (re)-installed the NICs and taking into account that the tulip driver is for PCI cards, I'd guess that there is a resource configuration problem, i.e. irq and i/o address. First check that the 'plug-and-play' operating system setting within the computer's bios is set to no or off. That way the hardware will configure p-n-p devices at boot time and linux should be happy. If its set to yes the latter will not happen because you are effectively saying that the operating system will do the resource allocation (as done in M$/Windoze).

Rgds

 

[marsd]

hi, TOM..
You have had some really good posts here in the past.. I
am surprised to see you asking for advice...
If serious then you need to check the mod names, even though you say the cards work after a manual reconfig. Depending on the distribution i would check the static config files for old settings that are loaded at boottime; this is the onlyreasonable answer to me. Why else would the cards fail on boot and then respond at the proper prep?

 

[MrTom]

hmmm, well i played around a bit last night and i think it has something to do with the tulip driver when two tulip cards are in at the same time.
i had two identical kingston cards in and replaced the (working) internal static IP card with an SMC card which uses a different module. on first boot the previously problematic kingston card got a DHCP lease straight away. this was with changing no settings at all, just swapping the one card out.
weird...

 

[ifincham]

Hi,

If you're still interested in getting both tulip cards to work you could try the diagnostic program from scyld which can configure the onboard eeproms --> ftp://ftp.scyld.com/pub/diag/tulip-diag.c .

You could also get debug messages from the driver by passing the debug=n parm (0 = nil 6 = everything) in /etc/modules.conf :

alias eth0 tulip
alias eth1 tulip
options tulip debug=3

And of course you can do 'cat /proc/pci' and 'cat /proc/interrupts' to get useful diagnostic information.

Rgds