Knute Johnson wrote:
I've got an FC6 box that I'm having some trouble getting the network
configured correctly. I keep getting the error below when eth0 is
started. This happens on reboot or if I attempt to restart the
network. Networking works but I keep getting the error message.
Any ideas where I've gone wrong?
Thanks,
knute...
[root@knute knute]# /sbin/service network restart
Shutting down interface eth0: [ OK ]
Shutting down loopback interface: [ OK ]
SIOCGIFFLAGS: No such device
Bringing up loopback interface: [ OK ]
Bringing up interface eth0: sysfs class device: Permission denied
Error, some other host already uses address 192.168.3.5.
[FAILED]
OK, since you and magnus are having the same issue, and at least one
other has confirmed a similar or same condition, we have enough to drop
the problem out of the anomaly real, and into something related to a
recent update (or something similar).
I have another piece of info that may or not be related:
The 2.6.20 kernels will NOT enable the ethernet driver for anaconda
properly.
I use:
linux text ks=http://server/ks/developer.ks
to begin an install. On three different systems now (all attempts I
have tried) this fails with interface failures. F3 shows DHCP READ
fails because network is DOWN, yet F4 shows the network up normally.
I roll my own initrd.img with the current kernels using pungi, and this
behavior began with 2.6.20 kernels.
The problems together indicate a larger issue with the basic ethernet
drivers in the 2.6.20 kernels.
So, the acid test is for you to boot an older kernel, check the problem,
boot the latest kernel, check the problem.
Is it a kernel (newer basic ethernet driver) problem?
Can you try a pre 2.6.20 kernel?
If it works, then we need to check bugzilla ...
