Have you experienced this? What's going on when this occurs? What do I need to do to keep it from occurring? Please advise. Thanks.
Dec 4 10:18:17 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:19 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:21 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:23 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:24 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:25 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:26 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:28 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:29 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:31 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:37 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:39 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:42 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:43 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:19:07 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:19:08 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:19:13 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:19:14 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:19:17 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:19:19 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:19:20 localhost kernel: e1000: eth1 NIC Link is Down
On Mon, Dec 27, 2010 at 05:22:38PM -0500, E Westphal wrote:
Have you experienced this? What's going on when this occurs? What do I need to do to keep it from occurring? Please advise. Thanks.
could be a bad network switch. or a bad cable on that NIC.
Though I wouldn't expect a bad cable to happen with such uniform periodicity (unless there's some physical disturbance that jiggles the cable every 4-5 seconds trigging an intermittent connection.)
have you tried just walking around the back of that box and watching the little blinky lights on the back to see if they go OFF every 4-5 seconds for a second or so then come back on? I'd expect either a bad cable or switch to cause them to go off, if they really have disconnected.
Dec 4 10:18:17 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:19 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:21 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:23 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:24 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:25 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:26 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:28 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:29 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:31 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:37 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:39 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:42 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:43 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:19:07 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:19:08 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:19:13 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:19:14 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:19:17 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:19:19 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:19:20 localhost kernel: e1000: eth1 NIC Link is Down
fred smith wrote:
On Mon, Dec 27, 2010 at 05:22:38PM -0500, E Westphal wrote:
Have you experienced this? What's going on when this occurs? What do I need to do to keep it from occurring? Please advise. Thanks.
could be a bad network switch. or a bad cable on that NIC.
<snip> I've got a bad feeling about this. You say this is eth1 - is eth0 in use? Is it doing similar things? We had a similar problem, and the NIC itself finally died. I think I managed to make it work a bit better for a few months by putting wooden wedges from the hardware store under the box, to let more air around to keep it cooler.
mark
On Mon, 2010-12-27 at 18:01 -0500, m.roth@5-cent.us wrote:
fred smith wrote:
On Mon, Dec 27, 2010 at 05:22:38PM -0500, E Westphal wrote:
Have you experienced this? What's going on when this occurs? What do I need to do to keep it from occurring? Please advise. Thanks.
could be a bad network switch. or a bad cable on that NIC.
<snip> I've got a bad feeling about this. You say this is eth1 - is eth0 in use? Is it doing similar things? We had a similar problem, and the NIC itself finally died. I think I managed to make it work a bit better for a few months by putting wooden wedges from the hardware store under the box, to let more air around to keep it cooler.
mark
Yea - eth0 works fine all the time. It's only eth1 that acts up. The last responder who said to look at a faulty solder joint has probably got it right. I've got a bud who is a gosh darn good EE and amateur radio builder - another one of us 'young' fellows who started out when 'core memory' was delivered by truck. Thanks all.
Ed
On Mon, Dec 27, 2010 at 5:22 PM, E Westphal enwestph@rochester.rr.com wrote:
Have you experienced this? What's going on when this occurs? What do I need to do to keep it from occurring? Please advise. Thanks.
Dec 4 10:18:17 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:19 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:21 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:23 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:24 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:25 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:26 localhost kernel: e1000: eth1 NIC Link is Down Dec 4 10:18:28 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:29 localhost kernel: e1000: eth1 NIC Link is Down
What is the NIC connected to? Have you tried a different switch port or network cable? Does rebooting the server make the problem go away for a bit? It seems like you need to perform some basic troubleshooting and let us know what you have an have no tried.
Ryan
On Mon, 2010-12-27 at 17:29 -0500, Ryan Wagoner wrote:
On Mon, Dec 27, 2010 at 5:22 PM, E Westphal enwestph@rochester.rr.com wrote:
Have you experienced this? What's going on when this occurs? What do I need to do to keep it from occurring? Please advise. Thanks.
Dec 4 10:18:28 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:29 localhost kernel: e1000: eth1 NIC Link is Down
What is the NIC connected to? Have you tried a different switch port or network cable? Does rebooting the server make the problem go away for a bit? It seems like you need to perform some basic troubleshooting and let us know what you have an have no tried.
Ryan _______________________________________________
I've tried all the obvious things. Cables, router ports, rebooting etc. It's quite strange. From a cold bootup, it's fine. Then about 45 minutes later it starts to act up. None of the obvious things seem to help. Then, about 90 minutes after bootup, everything is fine again! Rock solid! Ethtool and anything else I've tried fails to yield any hints, thus my question to all of you.
On Mon, 27 Dec 2010 17:37:50 -0500 E Westphal wrote:
I've tried all the obvious things. Cables, router ports, rebooting etc. It's quite strange. From a cold bootup, it's fine. Then about 45 minutes later it starts to act up. None of the obvious things seem to help. Then, about 90 minutes after bootup, everything is fine again! Rock solid! Ethtool and anything else I've tried fails to yield any hints, thus my question to all of you.
It sounds very much like a hardware issue, more specifically a cold solder joint whose properties change as the unit warms up.
On 12/27/2010 02:37 PM, E Westphal wrote:
On Mon, 2010-12-27 at 17:29 -0500, Ryan Wagoner wrote:
On Mon, Dec 27, 2010 at 5:22 PM, E Westphal <enwestph@rochester.rr.com mailto:enwestph@rochester.rr.com> wrote:
Have you experienced this? What's going on when this occurs? What do I need to do to keep it from occurring? Please advise. Thanks.
Dec 4 10:18:28 localhost kernel: e1000: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX Dec 4 10:18:29 localhost kernel: e1000: eth1 NIC Link is Down
What is the NIC connected to? Have you tried a different switch port or network cable? Does rebooting the server make the problem go away for a bit? It seems like you need to perform some basic troubleshooting and let us know what you have an have no tried.
Ryan _______________________________________________
I've tried all the obvious things. Cables, router ports, rebooting etc. It's quite strange. From a cold bootup, it's fine. Then about 45 minutes later it starts to act up. None of the obvious things seem to help. Then, about 90 minutes after bootup, everything is fine again! Rock solid! Ethtool and anything else I've tried fails to yield any hints, thus my question to all of you.
CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Are running the latest CentOS 5.5 release with all updates installed? Back around 5.0 through at least 5.1, possibly even 5.2, there were some driver bugs that affected some ethernet cards. I have also seen problems that showed up with cheap switch hardware and went away when plugged into a different switch. As far as inexpensive switches go, the low end HP managed switches are one of the few that I've had good luck with.
Nataraj
-
E Westphal -
Frank Cox -
fred smith -
m.roth＠5-cent.us -
Nataraj -
Ryan Wagoner