On 03/18/2013 12:57 PM, Stan Hoeppner wrote:
On 3/18/2013 3:57 AM, Casper Langemeijer wrote:
If I ifup both eth0 and eth1, ifconfig shows both interfaces. No
difference between output for them. (except ip/mac of course) The
counters get reset every few seconds for the broken eth0. Also: ethtool
shows nothing out off the ordinary except that it cannot detect Speed
and Duplex.
Do you see link flapping in dmesg, something like this, continuously?
....kernel: e100 0000:00:0d.0: eth0: NIC Link is Down
....kernel: e100 0000:00:0d.0: eth0: NIC Link is Up 100 Mbps Full Duplex
It's beginning to like you simply have a port that failed, and that
loading the firmware blob was coincidental.
Remember the old maxim: If ICs are going to fail in normal operation,
they typically do so within 72 hours. Which is why custom shops for
years advertised that they burn their servers in with real workloads for
72 hours (though few actually did). With the information I have so far,
I'd guess the transceiver for port0 on the card failed. It's a $0.25
IC. This would tend to explain why you can't get link/duplex status.
Every 5 seconds, the NIC is Down and Up again.
It gets even more weird: I had the NIC daughter card replaced, and the
problem still exists. I think I will use the machine in production as it
is now. Since I'm not using 4 ethernet ports, for me the issue is not
blocking. I hope someday someone encounters the same problem and finds a
solution.
I will receive a second server any day now. I will not be installing the
firmware-linux-nonfree package on that one...
Thank you for your time. If anyone else stumbles on this issue and wants
to contact me: Please do so off-list via personal email.
--
To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/51471565.9080...@langemeijer.eu
