2

Im using lenevo M5 server which have Gbps ethernet cards, but strangely when i read cat /sys/class/net/eth3/speed it show 100 similarly when using ethtool it show 100Mbps only.The output of dmesg and ethtool are below. Any idea why is that

root@server1:~# dmesg |grep eth1
[ 3.027487] tg3 0000:16:00.1 eth1: Tigon3 [partno(BCM95769) rev 5719001] (PCI Express) MAC address 08:94:ef:5d:81:a2
[ 3.027497] tg3 0000:16:00.1 eth1: attached PHY is 5719C (10/100/1000Base-T Ethernet) (WireSpeed[1], EEE[1])
[ 3.027503] tg3 0000:16:00.1 eth1: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[1] TSOcap[1]
[ 3.027508] tg3 0000:16:00.1 eth1: dma_rwctrl[00000001] dma_mask[64-bit]
[ 28.133071] bond0: Adding slave eth1
[ 28.249876] bond0: Enslaving eth1 as a backup interface with an up link
[ 53.149602] tg3 0000:16:00.1 eth1: Link is up at 100 Mbps, full duplex
[ 53.149620] tg3 0000:16:00.1 eth1: Flow control is off for TX and off for RX
[ 53.149624] tg3 0000:16:00.1 eth1: EEE is disabled

# ethtool eth3 

Settings for eth3: 
Supported ports: [ TP ] 
Supported link modes: 10baseT/Half 10baseT/Full 
100baseT/Half 100baseT/Full 
1000baseT/Half 1000baseT/Full 
Supported pause frame use: No 
Supports auto-negotiation: Yes 
Advertised link modes: 100baseT/Full 
Advertised pause frame use: Symmetric 
Advertised auto-negotiation: Yes 
Link partner advertised link modes: 100baseT/Half 100baseT/Full 
1000baseT/Full 
Link partner advertised pause frame use: No 
Link partner advertised auto-negotiation: Yes 
Speed: 100Mb/s 
Duplex: Full 
Port: Twisted Pair 
PHYAD: 2 
Transceiver: internal 
Auto-negotiation: on 
MDI-X: off 
Supports Wake-on: g 
Wake-on: g 
Current message level: 0x000000ff (255) 
drv probe link timer ifdown ifup rx_err tx_err 
Link detected: yes

When i try to change the speed with ethtool -s eth3 speed 1000 duplex full the "Speed " section is "Unknown" similarly reading(cat) the file /sys/class/net/eth3/speed gives invalid argument.

Improve this question
3
  • 4
    Just to be sure, is your cabling ok? It should be a cat 5e at least with all 8 wires patched.
    – Diamond
    Jul 31, 2018 at 10:24
  • @Diamant Cat6 cable is used to connect the server's ethernet port to switch. Do you know possible reasons for this behaviour ?
    – sherpaurgen
    Aug 3, 2018 at 12:45
  • 1
    The cable category is relative unimportant as Category-5e will work at gigabit. What is important is that you have all four pairs (eight wires) properly connected. Gigabit requires the use of all four pairs, but if it only has two good pairs, then it falls back to 100 Mbps. I have seen many cable installations and/or patch cables that only use two pairs because that was all that was needed for 10/100 Mbps, and people were trying to save money on cabling, Unfortunately, it does not work with 1000 Mbps.
    – Ron Maupin
    Aug 3, 2018 at 15:44

2 Answers 2

Reset to default
0

Check the port speed configuration for that specific port on the router, it could be limited to 100Mbps or the maybe the auto negotiation fails in which case you could fix it to 1000Mbps.

The commands depend on your router, here is a cisco example.

Improve this answer
2
  • Unfortunately i dont have access to the network switch but when i asked datacenter technicians about port speed the answer was it was 1000Mbps. Do you know if there are other possible reasons for this problem ?
    – sherpaurgen
    Aug 3, 2018 at 12:49
  • mm.. not really.. but it can't be set to a fixed 1000Mbps since you are getting 100. maybe let them configure the fixed setting? (no negotation) Or have them change network cables to be sure they're not faulty.. but these are just guesses of course
    – mx1up
    Aug 16, 2018 at 15:41
0

Thanks to @Diamant. After digging more the one of the CABLE connecting server eth ports seems to be faulty and replacing it solved the problem.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .