At Tue, 19 May 2009 09:04:43 -0400 CentOS mailing list centos@centos.org wrote:
I reimaged a compute node on our cluster with the latest 5.3 updates (we were previously running 5.2), but we kept the kernel at 2.6.18-92.1.10.el5 until I can find time to rebuild some of our kernel modules. After the image install finishes and the system reboots, the eth0 ethernet interface disappears. If I do an ifconfig a, I see what should be eth0, but it¹s listed as __tmp2081258173.
[root@node0770 ~]# ifconfig -a __tmp2081258173 Link encap:Ethernet HWaddr 00:1E:68:86:67:04 BROADCAST MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:0 (0.0 b) TX bytes:0 (0.0 b) Interrupt:66
The dmesg output isn¹t very helpful:
[root@node0770 ~]# dmesg|grep eth0 eth0: forcedeth.c: subsystem: 0108e:534b bound to 0000:00:08.0
If I remove our lustre modules that were built for the 2.6.18-92.1.10.el5 kernel and reboot, the eth0 interface reappears. Another piece to this puzzle is that this problem only seems to happen on our Sun X2200¹s. Our Dell 1950¹s work just fine after putting on the 5.3 updates. Anyone know what could cause this behavior?
Check /etc/modprobe.conf (and /etc/sysconfig/network-scripts/if-cfg-eth0) -- if you are doing a disk-to-disk backup type of install, the alias for eth0 is very likely wrong (and the HW address in /etc/sysconfig/network-scripts/if-cfg-eth0 is also wrong). You may have to manually update these two files on the 'new' machine, since it likely has a different NIC, requiring a different driver. It will also have a different MAC (HW) address as well. In the old days, kudzu would detect this and pop up during the boot process.
What does lspci display?
Thanks, Randy MIME-Version: 1.0
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos