After moving one of my servers to a new datacenter, my second gigabit NIC appeared to have stopped working.

ifconfig -a reveals that eth0 works fine, but there is now an inactive sit0 (ipv6-in-ipv4) sitting in the output. Now, while I was at the datacenter, I assumed sit0 was just something I can ignore and that eth1 was a dead port. Today, I logged into my switch to try and figure out which server's nic was plugged into which port (I forgot to trace), and I noticed that every server's NIC was plugged in AND reading as online (this included the server with eth0 and sit0).

It is my suspicion that the eth1 on the server is now reading as sit0. How exactly can I correct this? All of my servers have dual gigabit NIC's to support the internal and the local intranet (one nic for each). On the mission critical server, I am being forced to combine eth0 to act as both internet and LAN NIC. The problem is that it places over 40 Mbps of bandwidth onto that NIC at all times. Because of that, and the occasional load spike due to peak hours, I absolutely need to correct this situation.

Thank you.

sit stands for "simple internet transition" and is bassically a device
capable of encapsulating ipv6 in ipv4 datagrams.
Reference (http://www.ussg.iu.edu/hypermail/linux/kernel/9611.1/0330.html)

I think your problem is somewhere else.

That link suggests that I updated drivers before the move; I did not do that at all.

Eth0 and Eth1 work. I shut down the server. I move it to the new datacenter. I hooked it back up to the switch. I turn the server back on. Now sit0 shows up and eth1 does not. At the same time, eth1's activity lights are lit, and the switch says it is up.

Almost sounds like the kernel changed. Kernel updated? server shutdown, server installed in new DC, server comes up on another kernel not properly configured??

Just a guess, check your bootloader?

Also, have you tryed to bring eth1 up manually? Your not giving me much to work off of, so I have to make some general guesses.

Also, I don't think sit0 should be up, unless you specifically set it up.

ifconfig -a says eth0 and sit0 exist. ifup eth1 returns "eth1 does not exist."

sit0 should exist, but it should not be up.

What system is this on anyway? network links don't just drop off on their own, something happened somewhere.

This is a WhiteBox Enterprise Linux server. The only difference between point A and point B was the physical location. No software changed, no hardware changed, that all remained the same. What could have possibly happened to drop it from the system?

The only thing I can think of is a software change. I am assuming the machine was setup right at one point to survive a power cycle.

One thing that comes to mind, do you use yum for updates? Yum doesn't put the kernel in it's exclude= by default.

Also, whats the possibility that you or someone had to do some kernel work to get the second nic online? Maybe the right kernel isn't the default?

It isn't possible to turn it off and have it change, and very few things are dynamic on startup. Kernel is one.

Does either show up in your dmesg?

No software changed occuree, the same kernel is in use between boots, and we do not use Yum. dmesg is empty in terms to this problem.

Is this the new server that had that switch problem Jak? Could it be on the switch end? Also, you try playing around a removing sit0 and adding eth1?

Nah, the switch is fine, everything was working before the move. The port is working on the switch, and the everything is running smoother than ever.

What's the command to do that o_O; lspci acknowledges that two gigabit NIC's exist.