I was using a small box as a router and one of the ports started going off and coming back at 100M. I truly believe it is simply that it was a case of overheating. Although CPU temps were only about 67 degrees, the physical box itself was almost burning hot. I solved the cooling issue and never had the issue again.
Jul 28 15:09:27 swithbox kernel: e1000e: eth1 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Jul 28 15:09:28 swithbox kernel: e1000e: eth1 NIC Link is Down
Jul 28 15:09:30 swithbox kernel: e1000e: eth1 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: Rx/Tx
Jul 28 15:09:31 swithbox kernel: e1000e: eth1 NIC Link is Down
Jul 28 15:09:33 swithbox kernel: e1000e: eth1 NIC Link is Up 100 Mbps Full Duplex, Flow Control: Rx/Tx
Jul 28 15:09:33 swithbox kernel: e1000e 0000:02:00.0: eth1: 10/100 speed: disabling TSO
I just want to shed light that sometimes we may assume the port is just bad or the cable is bad, sometimes this can happen because of overheating. It has never happened again since I fixed the cooling where the CPU and physical box reached scorching temperatures that have never been seen.
On that note, it probably signals it is time to apply some new thermal paste onto the bridges and CPU, as it has probably dried out especially if your load is almost non-existent.
linux, nic, connecting, gigabit, speeds, overheatingi, router, ports, overheating, cpu, temps, cooling, jul, swithbox, kernel, eth, mbps, duplex, rx, tx, disabling, tso, scorching, temperatures, signals, thermal, bridges, dried, existent,