P. Thompson wrote:
OK similar question as before:
I think I've spotted a bug in ifconfig but net-tools does not seem to be on the list of supported apps for fedora in Bugzilla...
Basically a sub interface of eth0 (eth0:0) gets the counters from the unrelated eth1 interface despite being a different hardware device:
eth0:0 Link encap:Ethernet HWaddr 08:00:2B:C5:0D:7A inet addr:192.168.100.2 Bcast:192.168.100.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:1408372 errors:0 dropped:0 overruns:0 frame:0 TX packets:3783306 errors:0 dropped:0 overruns:0 carrier:0 collisions:67186 txqueuelen:1000 RX bytes:151553989 (144.5 Mb) TX bytes:1326172537 (1264.7 Mb) Interrupt:3 Base address:0x2c00
eth1 Link encap:Ethernet HWaddr 00:00:F8:22:7C:65 inet addr:192.168.1.1 Bcast:192.168.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:1408372 errors:0 dropped:0 overruns:0 frame:0 TX packets:3783306 errors:0 dropped:0 overruns:0 carrier:0 collisions:67186 txqueuelen:1000 RX bytes:151553989 (144.5 Mb) TX bytes:1326172537 (1264.7 Mb) Interrupt:3 Base address:0x4800
Is there an eth0 defined??
In recent code, the sub interface gets the counters from the primary interface and the sub-interface counters are included in the base device counters. This may be related to not having the eth0 interface active and avalaible to get its data from. If so, this is certainly a bug, but it may also be a configuration issue since in my experience there is almost always an eth0 defined and used before an eth0:1 gets defined.
I can think of another time where this might be of concern. I usually think of eth0:0 as equivalent to the base device eth0 and thus begin my aliases as eth0:1. I do not know if this is required, but maybe someone can answer that.
