How to know the network card speed

[hangyu]

we have two same model 100/1000 network card that plug into two different servers , but I found that their transfer speed is different , one is 2,xxx/minutes , another one is 1,1xx/minutes , I am not sure whether the setting have problem , can advise how do I check what speed (100 or 1000) is enabled ? thx

[Stuart Browne]

Use 'ethtool ', i.e. 'ethtool eth0'.

The output will be something similar to:

Settings for eth0:
Supported ports: [ MII ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Advertised auto-negotiation: Yes
Speed: 100Mb/s
Duplex: Full
Port: Twisted Pair
PHYAD: 1
Transceiver: internal
Auto-negotiation: on
Supports Wake-on: g
Wake-on: d
Current message level: 0x000000ff (255)
Link detected: yes


This is stating that it's a Gigabit-card running in 100Mb mode at full-duplex.

Given what you're saying, I'm picking that the auto-negotiation has picked half-duplex. Double check the switch ports to make sure they're at full duplex, then re-negotiate the conenction, or use ethtool to force full duplex.

[hangyu]

thx reply ,

the switch is set to support 1000bit/s

the below is the output of the command , server 1 ( slow ) seems the transfer speed of this server is slow while server 2 ( fast ) seems the transfer speed of this server is fast , can advise do I need to change any setting ? thx


server 1 ( slow )
---------------
Settings for eth0:
Supported ports: [ MII ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Advertised auto-negotiation: Yes
Speed: 1000Mb/s
Duplex: Full
Port: Twisted Pair
PHYAD: 1
Transceiver: internal
Auto-negotiation: on
Supports Wake-on: g
Wake-on: d
Current message level: 0x000000ff (255)
Link detected: yes

server 2 (fast)
---------------
Settings for eth0:
Supported ports: [ TP ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Full
Advertised auto-negotiation: Yes
Speed: 1000Mb/s
Duplex: Full
Port: Twisted Pair
PHYAD: 0
Transceiver: internal
Auto-negotiation: on
Supports Wake-on: umbg
Wake-on: d
Link detected: yes

[hangyu]

thx ,

I alo check with demsg , both servers have the same result

dmesg |grep eth0
tg3: eth0: Link is up at 1000 Mbps, full duplex.

[Hemmetter]

Hi hangyu,

can you give
$ ifconfig eth0

for both servers, are there errors/drops etc?


rgds
HGH

[hangyu]

thx reply ,

no drop and error found , both server is very very stable , but speed is acceptable but I just would like to know why the speed have great difference.


Server 1 (slow)
eth0 Link encap:Ethernet HWaddr 00:06:5B:F6:1F:0A
inet addr:192.168.0.1 Bcast:192.168.0.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:323113163 errors:0 dropped:0 overruns:0 frame:0
TX packets:445493740 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:3026881068 (2886.6 Mb) TX bytes:3954808375 (3771.5 Mb)
Interrupt:28

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:222959 errors:0 dropped:0 overruns:0 frame:0
TX packets:222959 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:30363458 (28.9 Mb) TX bytes:30363458 (28.9 Mb)


Server 1 (fast)
eth0 Link encap:Ethernet HWaddr 00:11:43:E6:11:42
inet addr:192.168.0.2 Bcast:192.168.0.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:77898326 errors:0 dropped:0 overruns:0 frame:0
TX packets:266921310 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:1509990853 (1440.0 Mb) TX bytes:1904061972 (1815.8 Mb)
Interrupt:48 Base address:0xecc0 Memory:dfbe0000-dfc00000

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:3931491 errors:0 dropped:0 overruns:0 frame:0
TX packets:3931491 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:275038573 (262.2 Mb) TX bytes:275038573 (262.2 Mb)

[Stuart Browne]

That all looks good.

What are you using to do the speed timing? Is it possible there's a routing issue between the devices you're doing the test with?

[Hemmetter]

Hi hangyu

There are lots of factors that may affect your transfer.

How differ the two systems?
Is the server hardware the same?
Due to the MAC-addresses of your NICs you have DELL hardware, right?

Do they have the same NICs and Firmware?
OS-Release? The same driver?
Are the systems connected to the same switch?

Are the systems under different load?
Do you r/w your transfer-data from/to disk?


rgds
HGH

[hangyu]

thx

How differ the two systems? they are RHEL3
Is the server hardware the same? the same server series
Due to the MAC-addresses of your NICs you have DELL hardware, right? Yes, they are dell

Do they have the same NICs and Firmware? they are same series server model , I think the NIC should be same model
OS-Release? The same driver? RHEL3 , build-in driver
Are the systems connected to the same switch? Yes

Are the systems under different load? , the loading is very low
Do you r/w your transfer-data from/to disk?
Yes


I would like to said , I have 5 Dell server ( similiar grade ) , only one server is very fast ( 2,xxx transfer rate ) , but all others are similiar speed ( around 1,xxx ) .

[hangyu]

I am not sure whether the server1's driver , setting etc is different so its transfer rate may up to 2,xxx bit /s , can advise how to check it ? eg . NIC driver ver , setting , fireware ver ? thx

[Hemmetter]

Hi hangyu,

check:
$ cat /proc/net/nicinfo/eth0.info

rgds
HGH

[Rob Leadbeater]

Hi,

Looking at your ifconfig output I see that the slow server is on IP address of 192.168.0.1.

There's nothing wrong with that, but that IP address is commonly used as a default gateway address. Maybe there's something else in your network that's misconfigured that's sending lots of traffic to the slow server...

Hope this helps,

Regards,

Rob

[IT_2007]

check with mii-tool command and see what is the speed on the network card.

[Bill McNAMARA_1]

I don't think mii-tool is accurate...
maybe there is an upgrade, but on 3.0ES systems I know for sure it only shows a max of 100MB connections even if you're running 1GB.

I still use mii-tool however as a quick method to check cable connectivity..

ethtool is the way forward (or via the /proc filesystem)

[David Kirk_2]

hangyu,

From the look of the ifconfig output, your slow server is transferring twice as much data as the fast one.

Server 1 (slow)
RX bytes:3026881068 (2886.6 Mb) TX bytes:3954808375 (3771.5 Mb)

Server 2 (fast)
RX bytes:1509990853 (1440.0 Mb) TX bytes:1904061972 (1815.8 Mb)

I guess that could be because the fast server was rebooted more recently than the slow server though. Or, it could be that the slow server is slow because it is just twice as busy as the fast one.

How are you measuring the speed of your links? What sort of load are the servers under when you do your testing?

I found a tool called Iperf (http://dast.nlanr.net/Projects/Iperf/ ) the other day. Give it a go, and let us know the results.

--
Later

David Kirk