BladeSystem Virtual Connect
Showing results for 
Search instead for 
Did you mean: 

BL680c G7 Redhat NIC enumeration issue

Trusted Contributor

BL680c G7 Redhat NIC enumeration issue

Florian had a NIC enumeration question:




Maybe someone here can give me some pointers to the following issue:


Checking via VC, the first NIC in our BL680c G7 server's profile is 00:17:A4:77:68:30


When booting up the server, the order of all eth* interfaces is mixed up , and not in the expected sequence.

Instead of eth0 = 00:17:A4:77:68:30 , we will find the MAC 00:17:A4:77:68:30 associated with eth8.


It's evident, that the order of NICs is not directly related to the (counting up) mac-addresses, but rather

the internal wiring of the ethernet chips / hardware tree.


The MAC 00:17:A4:77:68:30 is first in the profile's ethernet adapter list AND at the same time the first MAC

in the list of device bay information AND also, it is the 'lowest' of any of the MAC addresses, in terms of sorting.

So why wouldn't it be the eth0 interface then?


Is this intended in the way we see it here, or is there something wrong with our setup? Please advise.




Input from Vincent:




The interface ordering and numbering depends on the PCI enumeration done by the OS when it boots. It varies from one server model to another depending on the exact PCI architecture of the server, there’s even been instances of a ROM BIOS upgrade changing the order of the enumeration.

There’s probably nothing wrong with your setup, it just happens to be that way with the server model you’re working with.




Any other comments or suggestions?


Trusted Contributor

Re: BL680c G7 Redhat NIC enumeration issue

Input from Dave:




This article explains the ROM BIOS upgrade changes that Vincent mentions.

You might want to check to see if you have an RBSU setting to change enumeration to an earlier version and see if this fixes your network ordering.