On Mon, 2007-06-25 at 16:51 -0700, John R Pierce wrote:
> John Horne wrote:
> > Okay, that seems easy enough :-) Thanks.
> > However, is the anaconda NIC ordering the same as those listed by the
> > bios? For a mixed NIC server it could be important.
> >
>
>
> not always. I've got some Intel SE7501WV2 based dual xeon servers
> which have dual intel pro1000 NICs onboard... the BIOS, MS Windows,
> etc think the one labled '0' on the outside is in fact the first port,
> but RHEL2.1 and RHEL3 at least thought that they were swapped, and that
> eth0 was the port labeled '1', while eth1 is the port labeled '0'.
>
Hmm, well I guess I could use the bios ordering to set the first NIC
options, and if anaconda thinks anything different then just restart the
installation using whatever it (anaconda) thinks is the first NIC. A
pain, but basically it's either going to be 100Mb or 1000Mb! :-)
> I know I could probably force the issue, instead, I just live with it.
> those servers have been SO reliable I've never had to dink with them.
>
Once the installation has done, we tend to bond the interfaces together
(using active-backup mode). I have found that by creating simple udev
rules, the NIC ordering no longer changes over reboots. I can then
correctly set the NIC options in the ifcfg-ethx files.
John.
--
---------------------------------------------------------------
John Horne, University of Plymouth, UK Tel: +44 (0)1752 233914
E-mail: John.Horne@xxxxxxxxxxxxxx Fax: +44 (0)1752 233839
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
