You are not logged in.
I figure the best way to start this is with a little background. Firstly, I think the most important thing would be hardware.
I'm running a Gigabyte P45 based motherboard (EP45-UD3P) which has two Realtek PCIe NICs onboard (RTL8111/8168B) as well as an Intel Pro/1000PT (desktop edition, e1000e) in the first PCIe x1 slot. Other hardware include a Q6600, 2x2GB of generic DDR2, GeForce 8400GS, LG DVD writer (can't remember the model number off of the top of my head) and a Seagate 7200.11 1TB drive.
Now, the problem I have been having is that that network interfaces do not have the same name after the computer restarts. Generally, one of the Realtek (should be names eth1 and eth2) NICs will switch names with the Intel Pro/1000PT (eth0) on boot.
So, that's annoying, but not generally a large issue. I'm using net-rename (which I believe uses ifrename), and I setup the interfaces the way they should be in iftab. This requires me to take down the interfaces with ifconfic ("sudo ifconfig ethX down"), then running net-rename and then bringing the interfaces up.
Now an issue that just cropped up (and what caused me to post here) is that when I brought eth2 down (which is what the Intel NIC had decided to call itself this time) it seemed to arbitrarily change its name to eth3. Dmesg doesn't show anything about the interface going down, but does show a new interface called eth3 that looks identical to when the interface originally started up.
And now, after some fighting with it, it seems that networking is working properly, including doing the renaming of the interfaces, but I'm still suspicious of whether or not it will continue working properly later.
On another, possibly related note (because it could be indicative of a hardware issue, which I'm starting to suspect) is that sometimes on boot, kinit hangs when it starts mounting my roof FS (formatted XFS) readonly. And after I started having issues with the network and I rebooted, booting the CPUs failed (it tried all four cores) before (I assume) a kernel oops and a reboot. This happened three reboots in a row before things settled down.
The reason I rebooted in the first place was because my networking seemed to be operating very slowly. For example, logging in from my local network via SSH timed out quite a few times before I was able to reboot the system. It just seemed like eth0 (the Intel NIC) was being flakey.
Any suggestions/insight/tips are greatly appreciated, and thanks are given in advance. Thanks.
Offline