We recently noticed our 10gE NIC was reporting as devXXXXX. Does anyone know why this would be given this name as opposed to ethX?
Our System: The issue occurred on two separate servers. We are running RHEL4.9 & RHEL 4.5 which are on separate networks.
I know that I can change the name using nameif but it has to be run before the interface is up or else it fails.
This problem has been reported in a few cases on different systems. In some cases simply deleting the NICs entirely (both the interfaces and the hardware) from Kudzu seems to fix it after a reboot. In other cases the Kudzu database has simply become corrupt and needs to be reset. In that case, clearing out the modprobe aliases is key too.
The first process is described here:
http://www.linuxquestions.org/questions/linux-hardware-18/ethernet-card-dge-528t-and-kernel-2-6-a-350564/>
While a corrupt Kudzu database is discussed within the context of a Sun server running RHEL here (just do a ctrl-f on devXXX):
http://docs.oracle.com/cd/E19121-01/sf.x4600/819-4347-19/819-4347-19.pd
RedHat has also acknowledged the problem on their website:
https://access.redhat.com/knowledge/solutions/2956