[SOLVED] Testing speed of new dual NIC

C

charfix

Member
Dec 2, 2022
15
1
8
I just installed a PCI NIC with two 2.5 Gbps ports. I connected the two ports with a cat6 patch cable.
I would like to test whether the ports can actually receive/transmit at their advertised speed.
What I thought I should do is run iperf -s on one interface (enp3s0), and then use the other with iperf -c (enp4s0).
But it's not clear how to force packets to come out of a particular interface (enp4s0).
How might I achieve this?

I thought of assigning enp3s0 and enp4s0 static IPs, and creating a container hooked up to a new bridge interface with enp4s0, then running iperf from it.
But I must have misconfigured something because the container complained with I tried talking to the iperf sever.
 
Hi,

Good if we see the network configuration on your PVE server… However, I guess you can achieve this using the -B flag (see `man iperf`) on iperf e.g.:

Bash: 
iperf -s -B 10.0.0.1

On the client using the -c flag you can specify the interface to use including the -B option followed by the IP address of the interface e.g:

Bash: 
iperf -c 10.0.0.1 -B 10.0.0.2
 
/etc/network/interfaces
Code: 
auto lo
iface lo inet loopback

iface enp10s0 inet manual

auto enp3s0
iface enp3s0 inet dhcp

auto enp4s0
iface enp4s0 inet dhcp

auto vmbr0
iface vmbr0 inet manual
        address 192.168.2.112/24
        gateway 192.168.2.1
        bridge-ports enp10s0
        bridge-stp off
        bridge-fd 0
Meanwhile, I assigned enp3s0 and enp4s0 IPs manually with ifconfig enpXs0 192.168.4.X/24 so
ip a shows:

Code: 
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
2: enp3s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
    link/ether 88:c9:b3:b0:dd:f7 brd ff:ff:ff:ff:ff:ff
    inet 192.168.4.1/24 brd 192.168.4.255 scope global enp3s0
       valid_lft forever preferred_lft forever
3: enp4s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
    link/ether 88:c9:b3:b0:dd:f8 brd ff:ff:ff:ff:ff:ff
    inet 192.168.4.2/24 brd 192.168.4.255 scope global enp4s0
       valid_lft forever preferred_lft forever
    inet6 fe80::8ac9:b3ff:feb0:ddf8/64 scope link 
       valid_lft forever preferred_lft forever
4: enp10s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master vmbr0 state UP group default qlen 1000
    link/ether 50:e5:49:3d:ae:f3 brd ff:ff:ff:ff:ff:ff
5: vmbr0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether 50:e5:49:3d:ae:f3 brd ff:ff:ff:ff:ff:ff
    inet 192.168.2.112/24 scope global vmbr0
       valid_lft forever preferred_lft forever
6: veth103i0@if2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master fwbr103i0 state UP group default qlen 1000
    link/ether fe:95:88:e2:e2:de brd ff:ff:ff:ff:ff:ff link-netnsid 0
7: fwbr103i0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000
    link/ether ee:aa:39:d9:db:82 brd ff:ff:ff:ff:ff:ff
8: fwpr103p0@fwln103i0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master vmbr0 state UP group default qlen 1000
    link/ether 22:58:50:9e:ae:9b brd ff:ff:ff:ff:ff:ff
9: fwln103i0@fwpr103p0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue master fwbr103i0 state UP group default qlen 1000
    link/ether ee:fe:44:21:2b:64 brd ff:ff:ff:ff:ff:ff
10: tap100i0: <BROADCAST,MULTICAST,PROMISC,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master vmbr0 state UNKNOWN group default qlen 1000
    link/ether 26:c5:97:49:d2:8d brd ff:ff:ff:ff:ff:ff

I just tried
iperf3 -s -B 192.168.4.1 and iperf3 -c 192.168.4.1 -B 192.168.4.2 but it's showing ~40 Gbps which I think means the packets never actually go over the wire.

How do I force them to?
 
I found a ready-to-use solution on the web. Lightly modified, it is:

Code: 
#/bin/bash

#echo $1
#echo $2

# Use ip -d link list to see maxmtu value of interface
# maximum transmission unit (MTU) is a measurement representing the largest data packet that a network-connected device will accept.
#ip link set $1 mtu 9000
#ip link set $2 mtu 9000
# ----


#echo Thanks
ip netns add ns_server
ip netns add ns_client
ip link set $1 netns ns_server
ip link set $2 netns ns_client
ip netns exec ns_server ip addr add dev $1 192.168.10.1/24
ip netns exec ns_client ip addr add dev $2 192.168.10.2/24
ip netns exec ns_server ip link set dev $1 up
ip netns exec ns_client ip link set dev $2 up
ip netns exec ns_server iperf3 -s &
ip netns exec ns_client iperf3 -c 192.168.10.1

killall iperf3

ip netns exec ns_client iperf3 -s &
ip netns exec ns_server iperf3 -c 192.168.10.2

killall iperf3

ip netns del ns_server
ip netns del ns_client

With a high MTU I was able to confirm the NIC maxes out its claimed speed.
 
charfix said:
I found a ready-to-use solution on the web. Lightly modified, it is:

Code: 
#/bin/bash

#echo $1
#echo $2

# Use ip -d link list to see maxmtu value of interface
# maximum transmission unit (MTU) is a measurement representing the largest data packet that a network-connected device will accept.
#ip link set $1 mtu 9000
#ip link set $2 mtu 9000
# ----


#echo Thanks
ip netns add ns_server
ip netns add ns_client
ip link set $1 netns ns_server
ip link set $2 netns ns_client
ip netns exec ns_server ip addr add dev $1 192.168.10.1/24
ip netns exec ns_client ip addr add dev $2 192.168.10.2/24
ip netns exec ns_server ip link set dev $1 up
ip netns exec ns_client ip link set dev $2 up
ip netns exec ns_server iperf3 -s &
ip netns exec ns_client iperf3 -c 192.168.10.1

killall iperf3

ip netns exec ns_client iperf3 -s &
ip netns exec ns_server iperf3 -c 192.168.10.2

killall iperf3

ip netns del ns_server
ip netns del ns_client

With a high MTU I was able to confirm the NIC maxes out its claimed speed.
Click to expand...
Hey - thanks, Just wanted to let you know there's a v2 with this added as an arg, along with threads and time :) https://github.com/crazy-logic/iPerfCableTest
 
You must log in or register to reply here.
Top Bottom