On Mon, 2006-08-07 at 19:05 -0700, Blake Barnett wrote: > Ok, I have a slightly odd situation. I have multiple systems with > dead on-board NICs. We have added a new NIC to a PCI slot in each of > these machines and disabled the onboard NICs. The problem is, for > our automated deployment system to work smoothly, these new devices > _must_ be eth0 when the system comes up. The kernel always detects > the on-board NICs as eth0 and eth1, the new NIC that we've added > shows up as eth2. > > I know that one way around this is to use a netdev=<irq>,<dma>,eth0 > on the kernel append line, but unfortunately not all of these devices > have the same IRQ or memory address. I've seen various references to > disabling ACPI, and/or passing device driver-specific options. > Possibly also certain pci= options may help. I'm hoping someone has > dealt with this before.
On RH and FC, you just edit /etc/modrobe.conf and add the line: alias eth0 <modulename> That probably will do the trick for you. Michael > > Any ideas? > > Thanks, > -Blake > > /* > PLUG: http://plug.org, #utah on irc.freenode.net > Unsubscribe: http://plug.org/mailman/options/plug > Don't fear the penguin. > */ /* PLUG: http://plug.org, #utah on irc.freenode.net Unsubscribe: http://plug.org/mailman/options/plug Don't fear the penguin. */
