What is the best way to stop CentOS 6 from renaming the eth0 interface to em1? I've googled a bit and found many relevant posts, but have still not come up with an elegant solution. One of the resources I found was http://www.arachnoid.com/linux/network_names/index.html, and while the Python script contained therein does create a new udev rule that renames em1 back to eth0 it feels kludgy and requires me to manually create the ifcfg-eth0 script. I'm looking for a way to stop udev from renaming the interface in the first place, preferably something that allows me to automate it via kickstart so that any new systems I deploy will have a permanent fix. I'll push a fix to any systems I already have deployed with CentOS 6 that have this problem.
On some of my systems I have a 70-persistent-net.rules file with a comment that says it was generated by /lib/udev/write-net-rules, but I don't remember what I did to get it generate that rule, much less how to automate it and get it to work reliably (on one of my systems I have that rule but the interface is sill named em1). BTW, the em1 name only appears to be an issue on my newer Dell desktops; the older IBM/Lenovo systems I support are all using eth0
Thanks, Alfred