dhcp, multiple cards, no go !*#@
 

I a dual homed host I still have trouble bringing more than one interface up.

HW: PII, 3com + netgear + onboard 3com (boomerang)
Distro: SuSE 9.0 pro, updated via YOU

dhcp client: dhcpd (I have tried dhclient too, but that will only work with single interface).

In desperation I tried running one interface as static, but no joy.

Both console and YaST tried.

This is what bootmsg says when both are configured as dhcp clients:
Setting up network interfaces:

lo done
eth0 (DHCP) <notice>startproc: execve (/sbin/dhcpcd) [ /sbin/dhcpcd -d -D -N -t 999999 -l 120 -B -h ft-gw1 eth0 ], [ CONSOLE=/dev/console TERM=linux SHELL=/bin/sh OLDPWD=/etc/sysconfig/network progress=18 INIT_VERSION=sysvinit-2.82 RUN_FROM_RC=yes REDIRECT=/dev/tty1 COLUMNS=80 PATH=/sbin:/usr/sbin:/bin:/usr/bin:/etc/sysconfig/network/scripts RUNLEVEL=5 PWD=/etc/sysconfig/network PREVLEVEL=N LINES=25 HOME=/ SHLVL=4 splash=silent sscripts=40 _=/sbin/startproc DAEMON=/sbin/dhcpcd ]
dhcpcd[577]: broadcasting DHCP_REQUEST for 192.168.2.3


. . . . . . . . . . dhcpcd[577]: timed out waiting for DHCP_ACK response


dhcpcd[577]: broadcasting DHCP_DISCOVER


. . . . . no IP address yet... backgrounding.done
eth1 <notice>checkproc: /sbin/dhcpcd 577
(DHCP) <notice>startproc: execve (/sbin/dhcpcd) [ /sbin/dhcpcd -d -R -G -N -Y -t 999999 -l 120 -B -h ft-gw1 eth1 ], [ CONSOLE=/dev/console TERM=linux SHELL=/bin/sh OLDPWD=/etc/sysconfig/network progress=18 INIT_VERSION=sysvinit-2.82 RUN_FROM_RC=yes REDIRECT=/dev/tty1 COLUMNS=80 PATH=/sbin:/usr/sbin:/bin:/usr/bin:/etc/sysconfig/network/scripts RUNLEVEL=5 PWD=/etc/sysconfig/network PREVLEVEL=N LINES=25 HOME=/ SHLVL=4 splash=silent sscripts=40 _=/sbin/startproc DAEMON=/sbin/dhcpcd ]
dhcpcd[650]: broadcasting DHCP_REQUEST for 192.168.2.2


. . . . . . . . . dhcpcd[650]: timed out waiting for DHCP_ACK response


dhcpcd[650]: broadcasting DHCP_DISCOVER


. . . . . . no IP address yet... backgrounding.done




What to do/look for?
Your help much appreciated, I'm up against a blank wall.

you say
"dhcp client: dhcpd"

I think you might have just made a slip, because as your computer loads it says "dhcpcd" which is the correct daemon. DHCPd is the server, and if you are trying to get an address you want the client.

Moving along. Have you plugged another machine into this network and requested an IP address? Are both of these NIC's connected to the same network? or are they on seperate networks? Where is the DHCP server running on this network?

You are getting a timeout error. A firewall is blocking your DHCP request or there is no dhcp server running. Find the blockage.

You're right about the typo, it should read dhcpcd.

Regarding the blockage: How does the process run? I've tried to google for the process chain between host and client, but found nothing.

It seems to get a valid address (192.168.2.2 and .3), and then fails further down the process chain?

Further to the questions you raised above:
- I do get valid addresses and connetion with a laptop with PCMCIA eth card.

- I have tried connecting one card to the 192.168.2 network whilet the other was given a static IP on the 192.168.1 net. No joy.

- I have tried with both as dhcp clients on the 192.168.2 net, that resulted in the message log excerpt as above.

In all cases the dhcp server is my router, a Belkin wired/wireless one. (The wireless connections are not used in the above at all.)

- I have also tried to get the dhcp client to run against the DSL modem, which is my ultimate goal. No success.

What IP address is a valid one that you are talking about here? Maybe just paste me your ifconfig (or ipconfig /all if you are running windows).

Good job eliminating possible chances for error. Let's see what we can do with the wires.

So you want your dhcp client to get an address from the DSL modem? Or what? I don't get what you mean by "run against".

With a valid IP address I meant one in the range of the dhcp host, and seemingly issued by that host.

Ultimately I want the gateway to sit between the DSL modem and the Belkin router. In both cases I would prefer the gw to be a dhcp client on those two networks, but I can easily live with dchp on the DSL side and fixed IP on the router side.

And this is what the current ifconfig looks like:

eth0 Link encap:Ethernet HWaddr 00:01:02:9D:18:63
inet6 addr: fe80::201:2ff:fe9d:1863/64 Scope:Link
UP BROADCAST NOTRAILERS RUNNING MULTICAST MTU:1500 Metric:1
RX packets:59919 errors:0 dropped:0 overruns:0 frame:0
TX packets:4695 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
RX bytes:3595140 (3.4 Mb) TX bytes:2761618 (2.6 Mb)
Interrupt:10 Base address:0xdc00

eth1 Link encap:Ethernet HWaddr 00:C0:4F:95:A2:32
inet6 addr: fe80::2c0:4fff:fe95:a232/64 Scope:Link
UP BROADCAST NOTRAILERS MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:4683 errors:0 dropped:0 overruns:0 carrier:4683
collisions:0 txqueuelen:100
RX bytes:0 (0.0 b) TX bytes:2760898 (2.6 Mb)
Interrupt:11 Base address:0xdc80

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:95 errors:0 dropped:0 overruns:0 frame:0
TX packets:95 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:5994 (5.8 Kb) TX bytes:5994 (5.8 Kb)

ppp0 Link encap:Point-to-Point Protocol
inet addr:192.168.99.1 P-t-P:192.168.99.99 Mask:255.255.255.255
UP POINTOPOINT RUNNING NOARP MULTICAST MTU:1492 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:1 errors:0 dropped:27 overruns:0 carrier:0
collisions:0 txqueuelen:3
RX bytes:0 (0.0 b) TX bytes:60 (60.0 b)

I'm running out of ideas here. You say that the laptop got an ip that was valid, I dont recall you telling me which ip that was, but you told me which block it belonged to. In a last hurrah I would pull one of the cards out of the linux box. Treat it like you would your laptop for now, just a regular machine. Then work up. I have no idea what it could be, except that for some reason the cards are not connecting to the dhcp servers. Start again at square 2, and work your way up. If this whole backtracking thing doesn't work, there's always square 1 which is a reinstall.

Am I the one with a twisted mind ? A router (gateway) with dynamic IPs seems such an unlikely idea...

See, on each side of a router, machines are to be given routes indicating which traffic must go through which gateway. For example the default gateway is the way that should take any traffic which is not regarded as internal (that is, destination network address given by the n first bits of source IP, n being the number of bits set at 1 in the subnet mask); this default gateway is either hard-set in the hosts network settings, or pushed by the dhcp server.

Now the dhcp server does not know what a network adapter is used for (workstation or router), thus the default gateway (for instance) must be hard set in the dhcp server config, so obviously the router *cannot* be a dhcp client...

Perhaps this won't tell you why you have trouble configuring dhcp client on your gateway interfaces, but at least it might tell you why you shouldn't ;)

But, as I said, perhaps I am the one who is all wrong...

The dhcp client in gw and then server in the router is a point-to-point connection, so not much trouble.

Anyhow, I've given in, and use a "hw-firewall" and then the proxy in the internal net. It works, so I'll leave it like that for now.

do you need dhcpd to be broadcasting on both your network interfaces? probably not. you can specify what interfaces you want to use as an argument to dhcpd (or in your dhcpd startup script, whichever you want)

The network being small I have done as recommended in Ibsen's Peer Gynt: I've circumvented the problem.

I have put the proxy as a dhcp client with IP-address on a permanent lease. It does the job, although strictly speaking it is a somewhat untidy solution.