speed/duplex issue

cisco24x7 · ‎02-23-2008

Can someone explain this:

Cisco catalyst 2960 shows interface as 100-full, but the Redhat Linux ES

version 3.0 box connected to the switchport G0/10 is shown as

100-half. Which one is the correct setting then? Everything is working

fine with NO CRC errors.

C2960#sh mac address-table int

C2960#sh mac address-table interface g0/10

Mac Address Table

-------------------------------------------

Vlan Mac Address Type Ports

---- ----------- -------- -----

2 00a0.c90d.2a1d DYNAMIC Gi0/10

Total Mac Addresses for this criterion: 1

C2960#sh int g0/10

GigabitEthernet0/10 is up, line protocol is up (connected)

Hardware is Gigabit Ethernet, address is 0019.551b.d60a (bia 0019.551b.d60a)

Description: NGx R65 GW1 eth0

MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec,

reliability 255/255, txload 1/255, rxload 1/255

Encapsulation ARPA, loopback not set

Keepalive set (10 sec)

Full-duplex, 100Mb/s, media type is 10/100/1000BaseTX

input flow-control is off, output flow-control is unsupported

ARP type: ARPA, ARP Timeout 04:00:00

Last input never, output 00:00:09, output hang never

Last clearing of "show interface" counters 00:00:58

Input queue: 0/75/0/0 (size/max/drops/flushes); Total output drops: 0

Queueing strategy: fifo

Output queue: 0/40 (size/max)

5 minute input rate 19000 bits/sec, 22 packets/sec

5 minute output rate 163000 bits/sec, 105 packets/sec

5654 packets input, 550697 bytes, 0 no buffer

Received 160 broadcasts (0 multicast)

0 runts, 0 giants, 0 throttles

0 input errors, 0 CRC, 0 frame, 0 overrun, 0 ignored

0 watchdog, 157 multicast, 0 pause input

0 input packets with dribble condition detected

10428 packets output, 5360846 bytes, 0 underruns

0 output errors, 0 collisions, 0 interface resets

0 babbles, 0 late collision, 0 deferred

0 lost carrier, 0 no carrier, 0 PAUSE output

0 output buffer failures, 0 output buffers swapped out

C2960#

[Expert@NGx-gw1]# ethtool eth0

Settings for eth0:

Supported ports: [ TP MII ]

Supported link modes: 10baseT/Half 10baseT/Full

100baseT/Half 100baseT/Full

Supports auto-negotiation: Yes

Advertised link modes: 10baseT/Half 10baseT/Full

100baseT/Half 100baseT/Full

Advertised auto-negotiation: No

Speed: 100Mb/s

Duplex: Half

Port: Twisted Pair

PHYAD: 1

Transceiver: internal

Auto-negotiation: off

Supports Wake-on: d

Wake-on: d

Link detected: yes

[Expert@NGx-gw1]#

[Expert@NGx-gw1]# ifconfig eth0

eth0 Link encap:Ethernet HWaddr 00:A0:C9:0D:2A:1D

inet addr:192.168.1.201 Bcast:192.168.1.255 Mask:255.255.255.192

UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1

RX packets:12982488 errors:0 dropped:0 overruns:0 frame:0

TX packets:3284111 errors:0 dropped:0 overruns:0 carrier:0

collisions:0 txqueuelen:1000

RX bytes:3363084973 (3207.2 Mb) TX bytes:1925104265 (1835.9 Mb)

Interrupt:11 Base address:0xdec0 Memory:f8001000-f8001038

[Expert@NGx-gw1]#

aijaz802 · ‎02-23-2008

Hi,

Please re-check that the cable going from the Linux box is directly connected to the Gig0/10 of switch without any intermediate hub/switch.

Regards,

*aijaz*

cisco24x7 · ‎02-24-2008

The linux box is directly connected to the

Catalyst 2960 G0/10:

C2960#sh run int g0/10

Building configuration...

Current configuration : 141 bytes

!

interface GigabitEthernet0/10

description NGx R65 GW1 eth0

switchport access vlan 2

speed 100

duplex full

spanning-tree portfast

end

C2960#

The cable is also a good cable as well CAT 5+.

This cable works with other Cisco devices

without issues.

Any other ideas? Thanks.

CCIE Security

mohammedmahmoud · ‎02-24-2008

Hi,

I think that this is related to auto-negotiation, as appears from your output, your linux box is configured with auto-negotiation. As a best practice both sides must be configured with auto-negotiation, or both should be configured with the same manual speed/duplex, since most of the devices return to half duplex if it is configured with auto-negotiation and the other end is manually configured.

BR,

Mohammed Mahmoud.

cisco24x7 · ‎02-24-2008

"I think that this is related to auto-negotiation, as appears from your output, your linux box is configured with auto-negotiation"

Huh? The interface is EXPLICITLY configured

for 100-full. See below:

[Expert@NGx-gw1]# /sbin/ethtool -s eth0 speed 100 duplex full autoneg off

[Expert@NGx-gw1]# /sbin/ethtool eth0

Settings for eth0:

Supported ports: [ TP MII ]

Supported link modes: 10baseT/Half 10baseT/Full

100baseT/Half 100baseT/Full

Supports auto-negotiation: Yes

Advertised link modes: 10baseT/Half 10baseT/Full

100baseT/Half 100baseT/Full

Advertised auto-negotiation: No

Speed: 100Mb/s

Duplex: Full

Port: Twisted Pair

PHYAD: 1

Transceiver: internal

Auto-negotiation: off

Supports Wake-on: d

Wake-on: d

Link detected: yes

[Expert@NGx-gw1]#

aijaz802 · ‎02-24-2008

Hi,

I think Mohammed is right...

It seems that the linux box is set to manual with 100 MBPS n half....its NOT negotiating(its capable) as shown.. "Advertised auto-negotiation: No ". Try to set it to auto or make it 100-full.

Regards...

cisco24x7 · ‎02-24-2008

[Expert@NGx-gw1]# /sbin/ethtool eth0

Settings for eth0:

Supported ports: [ TP MII ]

Supported link modes: 10baseT/Half 10baseT/Full

100baseT/Half 100baseT/Full

Supports auto-negotiation: Yes

Advertised link modes: 10baseT/Half 10baseT/Full

100baseT/Half 100baseT/Full

Advertised auto-negotiation: No

Speed: 100Mb/s

Duplex: Half

Port: Twisted Pair

PHYAD: 1

Transceiver: internal

Auto-negotiation: off

Supports Wake-on: d

Wake-on: d

Link detected: yes

[Expert@NGx-gw1]#

It stays up a "full" for about 10 seconds, then goes back to "half". Weird part is

that there are no collisions on the catalyst interface. Very strange.

glen.grant · ‎02-24-2008

You probably won't see any errors on the Catalyst because that side is hardcoded to 100/full . If you want to leave the 2960 as is hardcode the nic card to 100/full . They must match on both sides . You may not see a whole lot errors until you start pumping larger amounts of traffic down that pipe and then you will really notice the performance degradation .

Danilo Dy · ‎02-24-2008

Hi,

It is strange! But you just clear the counters 58 seconds ago.

In one implementation, I have a 3750 GE interface hardcoded 100Mbps/Full-duplex and the HP-Unix System Engineer forgot to hardcode the server NIC and left it in Auto/Auto, the HP-Unix NIC negotiated the link as 100Mbps/Half-duplex. The collission registered in the switch is high and accumulating very fast, on top of that the data transfer is very slow.

If you can't make the Linux NIC to 100Mbps/Full-duplex, replace/recrimp the cable and make sure its straight through cable.

BTW, just to confirm the following about the collission error, please check;

1. The interface in the switch is the correct interface where the Linux is connected

2. Make sure that the Linux is not using NIC-Teaming and the other NIC is currently active

3. If the Linux are in HA, the other box might be the active.

Regards,

Dandy

cisco24x7 · ‎02-24-2008

1- Linux box is NOT NIC-Teaming. There is

ONLY 1 NIC in the box,

2- I hard code the Linux box to 100/Full with

this command:

[Expert@NGx-gw1]# /sbin/ethtool -s eth0 speed 100 duplex full autoneg off

However, the interface is still shown as

half. The catlyst is shown as 100/Full.

3- If this is true, I should get a lot of

errors on the interface, but I do not. I

am pumping about 90mpbs of ssh traffic

through this interface for 8 straight hours.

What makes it strange is that both the

Linux box and the Catalyst shows NO errors.