Device eth0 does not seem to be present, delaying initialization.
 

Okay, I have a really old, netpliance sort of distro that was installed onto a really old machine, built on a really old version of CentOS, and recently it became necessary to migrate the hard drive out to a newer and better machine. It's been like pulling teeth obtaining a machine I could get this to work in, but I'm most of the way to getting it up and running again. My problem is, on the newest machine, when I boot the thing, I'm getting the following error:

Tulip device eth0 does not seem to be present, delaying initialization.

....and ethernet is not working. The big problem is, this box is one of those configure-and-command-it-from-the-built-in-web-interface things, so with the ethernet part of it not working... that kinda puts a damper on using the thing.

Several of the general-Linux-help pages I've come across telling me what to do when I get an error like this tell me to go to go remove /etc/udev/rules.d/70-persistent-net.rules and then reboot, and it will then recreate that with the correct ethernet device inserted. Trouble is, when I invoke RM /etc/udev/rules.d/70-persistent-net.rules it tells me that this directory or file does not exist.

Some have suggested the kernel doesn't recognize or support the particular make of ethernet port on the machine, but I'm guessing its also possible its simply tripping over some internal setting that's not letting it "see" that the ethernet port is different from what it had on the previous machine. In any event, the current machine is an HP Compaq 8200 Elite CMT, which appears to date from about when Windows 7 came out, and I'm trying to use CentOS 4.9 on it.

(I need to do a complete backup of its info, from its internal functions for that, and then update this thing through at least two major-update versions of the netpliance system, but I ran into a severe ram shortfall in the original machine, and it seems pretty likely this was preventing the backup from even starting, hence the drive migration.)

When I invoke ifconfig -a I see only bond0 as my ethernet, with HWaddr 00:00:00:00:00:00

Anyway, I'm at a loss as to what to do.

Network device hardware details might be useful here...

That just yields me invalid option -- k

...and then lists a bunch of switches I can invoke. Was the command meant to list to me some info on how my system is configured?

That is due to the old version of CentOS I guess. Try
We're looking for output pertaining to your NIC, specifically the chipset details and if any driver is loaded.

I think the output is too long for me to see all of it. Here's the screen-grab of it after the command finished.

That really doesn't add anything. How about the output of
Sounds like you need to make a modern install. How old is this really old system? Are you going from/to the same architecture? Same processor? If not then it may not work at all.

If you need to catch that to text file then

Yes, redirect the output to a text file, and save to a memory stick so that you can post via an internet-connected PC.

Okay, how do I do that? What commands do I use, I mean.

Tekk showed an example of that already. As you mention that your old version of CentOS uses a version of lspci not supporting the '-k' switch, you could run the command as I already gave it and redirect the output to a text file. It's not clear to me what level your Linux knowledge is at, but here's what you could do...
Hopefully, you know how to mount a USB stick filesystem, copy said file to that, and unmount it again?

This should be applicable to CentOS 4....
https://www.slothparadise.com/mount-...rive-centos-7/

With USB sticks, about all I know is: plug it into the USB port. I wouldn't know what name to look for it under in the command line interface, or what commands to issue to mount or unmount it.

In the mean time look at the output of the command
You should see something like
The important information we need to know is what it outputs for your device:
Also it would help to the output of the command
uname -a

Okay, lspci -nnv | grep Ethernet yielded nothing, it just dropped back to the command line without reporting anything.

And uname -a resolved:

Linux server 2.6.9-103.ELsmp #1 SMP Fri Dec 9 04:31:51 EST 2011 i686 i686 i386 GNU/Linux

As suggested the older kernel does not recognize the Ethernet adapter which is why the output was blank.

You can also slowly browse through the output of lspci via less to confirm the above.

lspci | less

The kernel version does match the CentOS version.

According to HP specifications the 8200 Ethernet adapter is
The 82570LM uses the e1000e which is the same as mine as posted above. The Intel Pro 1000 CT uses the e1000 driver. You could download the driver source code and if you had a suitable environment compile it yourself...

Fraid I don't have enough experience in Linux stuff to compile it. (Or is that just a matter of plopping the source code in there somewhere and telling it to compile it? 0o) Is it possible someone else somewhere already did that and placed it somewhere already ready already?

Would simply placing a different, still-being-sold-but-fairly-old-model pci ethernet card in there get it recognized and enabled?