Okay, I am a little stumped by this and am mostly asking what the best way to proceed might be with this. CentOS 6 KVM guest running on a CentOS 6 KVM host. For various reasons I needed to restore this virtual machine from a backup. Restore seemed to behave perfectly well, and the client BOOTS after the restore - but the networks fail. Apparently (since I can get in on the console) the issue here is that it's trying to create the network interfaces on eth0 and eth1, which is how the server was previously. But after the restore the client only seems to have devices for eth2 and eth3 --- so it cannot start up the devices. The network config scripts are for eth0 and eth1.
Having this machine up today is not crucial to my life but -- how would you proceed with fixing this while causing the least long-term carnage? I don't understand exactly what happened here.
Comments
Submitted by cruiskeen on Sat, 04/15/2017 - 13:27 Comment #1
It appears that as the system is booting, udev is renaming the devices --
udev renamed eth0 to eth2 udev renamed eth1 to eth3
Submitted by cruiskeen on Sat, 04/15/2017 - 14:02 Comment #2
removed the network rules in /etc/udev/rules.d/70-persistent-net.rules and explicitly set the HWADDR in the network startup scripts, and now it starts and comes up on the network.
Submitted by JamieCameron on Sat, 04/15/2017 - 21:51 Comment #3
I suspect the cause here is that the MAC address changed between the original system and the one restored, which causes CentOS to pick different ethN names. Do you have any way to check if that was the case?