ethX ports playing musical chairs

Reindl Harald h.reindl at thelounge.net
Sat Feb 9 18:39:08 UTC 2013



Am 09.02.2013 19:30, schrieb Sam Varshavchik:
> Reindl Harald writes:
> 
>>
>>
>> Am 09.02.2013 19:07, schrieb Sam Varshavchik:
>> > Occasionally, upon reboot, apparently my eth0 and eth1 end up getting "swapped". Hillary ensues because HWADDR in
>> > /etc/sysconfig/network-scripts/ifcfg-Auto_eth[01] no longer matches:
>> >
>> > Feb  9 12:53:46 shorty network[1325]: Bringing up interface Auto_eth0:  ERROR    :
>> > [/etc/sysconfig/network-scripts/ifup-eth] Device eth0 has different MAC address than expected, ignoring.
>> >
>> > etc…
>> >
>> > I'm tired of logging in on a console, and initiating another reboot. eth0 and eth1 are different hardware. One is
>> > handled by tha atl1c kernel module, the other port is handled by the r8169 kernel module.
>> >
>> > Anyone happen to know how to get the drivers to come up in stable order?
>>
>> first throw away ANY MAC-reference from ifcfg*-files
>> nobody needs them because "DEVICE=ethx" is pretty clear
>>
>> second "/etc/udev/rules.d/70-persistent-net.rules" still works
>> on 999 out of 1000 machines
> 
> Well, apparently I'm the 1 in 1000, because mine already has these rules, the timestamp on that file is almost two
> years old, but occasionally the kernel still boots with the ports out of order. Pulling the MAC addresses out of
> the ifcfg files would probably mean that the public IP address would come up on my LAN port, and my Internet port
> get 192.168.0.1, which isn't really better than not having those ports come up at all

another option would be get rid of the "eth*" at all

* make "lan0" and "wan0" in udev-rules
* rename the ifcfg-scripts
* modify "DEVICE=lan0" and "DEVICE=wan0" in the ifcfg-scripts

from this moment on kernel-timig should not matter at all

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 263 bytes
Desc: OpenPGP digital signature
URL: <http://lists.fedoraproject.org/pipermail/users/attachments/20130209/6404b862/attachment.sig>


More information about the users mailing list