Solved: Frustrated with Network Devices Setup

So I decided to try openSUSE, after reading over and over on the Kubuntu forums that is was so much better. Over all the experience has been good, with this one rather large blemish.

I have been trying for about a week to set up a home server, with the minimal server set up. After getting the minimal server installed, I moved on to setting up a static IP, which openSUSE just doesn’t seem to like. After setting up the IP using YaST2 and saving my settings, I would notice I don’t have a network connection. So I would go back into the Network Device set up and find the following.

Name                               |IP Address
CK804 Ethernet Controller          |Not Configured
88E8001 Gigabit Ethernet Controller|Not Configured
CK804 Ethernet Controller          |192.168.0.10

Now CK804 Ethernet Controller is the one I set up, when it was eth0, then it duplicates itself, removes the MAC Address and disables itself.

Occasionally I will get it to work, and get out and be able to establish a connection, and even SSH into the server, but god help me if I ever have to reboot the machine because on reboot it just kills the configuration and duplicates again.

So my question is, do I trash openSUSE and go back to Kubuntu or is there something I can do to fix this, because hours, DAYS of researching has rewarded me with nothing but a headache.

What is the definition of Stupidity? Doing the same thing over and over, expecting a difference response.

Thanks in advance for any help in this,
David.

Not my Forte but did you check John’s page
Linux HOWTOs and Tutorials: Suse Linux 10.0, 10.1 openSUSE 10.2, 10.3, 11.0, 11.1

Yep, maybe I should add the tutorials I have tried.

I started with this Tutorial

The Perfect Server - OpenSUSE 11.1

After the issues, I moved to Swerdna’s tutorial and even a tutorial on building a Home Server, which I abandoned after realizing that 75 percent of that tutorial was pretty pictures which the author either forgot to embed or which the source for the embedding is no longer available, leaving me with criptic instructions and no explanation for the why of what I was doing. Which is also sort of the point, to learn while I do.

Thanks for the fast reply by the way :slight_smile:

I have never encountered a problem like this.

I suspect the driver for CK804.
Can you configure the other card (88E8001) and try? This is to verify that the culprit is CK804 or not.

Also, please have a look at the files in /etc/sysconfig/network directory. You should see files like ifcfg-eth0, if-cfg-eth1 etc. where configuration is save permanently.

Which board are you using?
Are you sure the static IP isn’t being used by another device on your network?
Consider using the Marvell NIC instead of the NVidia NIC.

Uwe

I’ve not attempted to configure the other nic, mostly because my ultimate goal is to configure this box as a gateway. Here are some more things going on.

In /etc/sysconfig/network I found, as you would expect, multiple ifcfg-* files.

I decided to check /etc/udev/rules.d/70-persistent-net.rules and found a list of about 10 rules with multiple eth* devices and varying mac addresses.

At this point I cleaned up the /etc/sysconfig folder, by taking out the erroneous ifcfg- files, then I ran rm /etc/udev/rules.d/70-persistent-net.rules and rebooted to start with a clean slate with which to work on.

When I rebooted, I reconfigured the network adapter with the same IP address and I checked my settings against my workstation box. Everything was clean, I finished out of the set up, and let it run the configuration. Once at CLI, I ran a ping Google and verified that I had a connection. I then attempted to reboot. Upon restart eth0 and eth1 were both unconfigured and eth4, which I didn’t set up, was not working. I then went back into the network set up again and I was staring at

Name                               |IP Address
CK804 Ethernet Controller          |Not Configured
88E8001 Gigabit Ethernet Controller|Not Configured
CK804 Ethernet Controller          |192.168.0.10

and there was once again no MAC address on the duplicate controller.

I am using a DFI Lanparty NForc4-SLI-Expert board. Both NICs work under ubuntu, without this sort of strangeness, but I am willing to try with the Marvell Controller. I will admit though that if I am unable to get both NICs working in openSUSE, then this is an exercise in futility as I will need both for my intended configuration.

I am at my wits end on this. I have never encountered anything like this before.

I did think I had it configured till it hit the point in the reboot that it started counting down the 30 seconds as it attempts to initialize the NIC. Could it be that I need to do a shutdown and not a reboot?

Again, thanks for the replies and the suggestions,
David

Here’s the update

If you are drinking men/women join me in a drink, this is one for the record books. I have owned this board for years, and this is the first time I have encountered this with static IPs off the nvidia controller. It seems that the controller doesn’t have a burned in MAC address. I thought it did. I knew you could go in and alter the MAC, but that smacks of having something to hide, so I never did it. Every time I rebooted openSUSE it assigned a new MAC address to the controller there by changing the controller. This was causing all the chaos.

Thank you, buckesfeld and syampillai, I was so intent on bending the installation to my will, that I forgot to test against the other controller. In my field, testing the alternative is crucial and yet I forgot that simple little rule. This was a good lesson on many levels.

I would like to say, it’s a testimony to OpenSUSE that I stayed with the problem this long. I had just thrown in the towel when my blackberry buzzed saying I had mail.

Thanks for your help, and thanks for believing in this distro enough to help me get through this tribulation.

David

Hmmm, is there a way to add resolved to the thread?

Hey man, this is truely great: getting back at the forum to explain your mistakes.

Maybe you could post an idea about adding “Solved” to threads that do contain a solution to the problem mentioned in the thread title

Well, those crappy NVidia onboard NICs (using kernel module “forcedeth”, what a matching name ;-)) are known to do such utterly stupid things.

However, to resolve this “every boot gives a new interface with other MAC”-chaos, there is a rather simple solution.

  • Go to YaST => Network Devices => Network Settings

  • Choose card and Edit => Hardware => udev rules

  • Change from “MAC-Address” to “Bus ID”.

If you rely on the card always having the same MAC-address (i.e. for DHCP) you can also use the “LLADDR=”-option in the respective ifcfg-ethX file.

Read /etc/sysconfig/network/ifcfg.template for more information.

This is a good thread that solves a common problem with the driver.
Well done!

I’ll add “solved” to the title if there are no objections.
[edit]
Someone beat me to it :slight_smile:
[/edit]

Good to hear that the problem is identified. Everyday, I learn something new.

This was an amazing experience all the way around, admittedly I almost gave up on it, but this is how you learn things. There is no way sitting in a classroom could have prepared me for something like this.