Configuration to build iperf3 images
This repo and its associated public image make it easier to run iperf3 networking tests in containerized environments.
The built image is available at:
docker pull stmassey/iperf3-server
Details on the iperf3-client image can be found at https://github.com/ibm-messaging/iperf3-client
You can run the client and server on the default docker bridge network (or even one of your own defined networks), but its more likely you will want to run it on the host network to check network performance between different servers
docker run -itd --net="host" stmassey/iperf3-server
You can check the logs to see which ip addresses are available to connect to the iperf3 server. Select one of those and pass it to the client as an envvar:
docker run -it --net="host" --env IPERF3_HOSTNAME=<hostname> stmassey/iperf3-client
iperf3 testing: client
Running against host: <hostname>
Connecting to host <hostname>, port 5201
[ 5] local *.*.*.* port 12904 connected to *.*.*.* port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 2.55 GBytes 21.9 Gbits/sec 0 895 KBytes
[ 5] 1.00-2.00 sec 2.32 GBytes 19.9 Gbits/sec 0 895 KBytes
[ 5] 2.00-3.00 sec 2.32 GBytes 19.9 Gbits/sec 0 1.37 MBytes
[ 5] 3.00-4.00 sec 2.32 GBytes 19.9 Gbits/sec 0 1.37 MBytes
[ 5] 4.00-5.00 sec 2.55 GBytes 21.9 Gbits/sec 0 1.37 MBytes
[ 5] 5.00-6.00 sec 2.37 GBytes 20.3 Gbits/sec 0 1.37 MBytes
[ 5] 6.00-7.00 sec 2.32 GBytes 19.9 Gbits/sec 0 1.37 MBytes
[ 5] 7.00-8.00 sec 2.32 GBytes 19.9 Gbits/sec 0 1.37 MBytes
[ 5] 8.00-9.00 sec 2.32 GBytes 19.9 Gbits/sec 0 1.37 MBytes
[ 5] 9.00-10.00 sec 2.33 GBytes 20.0 Gbits/sec 0 1.37 MBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 23.7 GBytes 20.4 Gbits/sec 0 sender
[ 5] 0.00-10.04 sec 23.7 GBytes 20.3 Gbits/sec receiver
If you want to run it on the docker network, then the following commands might be useful:
docker run -itd stmassey/iperf3-server
docker run -it --env IPERF3_HOSTNAME=$(docker inspect --format "{{ .NetworkSettings.IPAddress }}" $(docker ps -ql)) stmassey/iperf3-client
The easiest way to launch the iperf3-server in an OCP environment is to create a new workspace and add the server project as a new application:
oc new-project iperf
oc new-app stmassey/iperf3-server
This will create the required Kube artefacts (deployments, services, pods). You can check the available interfaces either by examining the pod logs:
oc logs pod/iperf3-server-6997546b96-p2p57
iperf3 testing: server
Local IP Addresses:
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
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
3: eth0@if1689: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1450 qdisc noqueue state UP group default
link/ether 0a:58:0a:82:04:c8 brd ff:ff:ff:ff:ff:ff link-netnsid 0
inet 10.130.4.200/23 brd 10.130.5.255 scope global eth0
valid_lft forever preferred_lft forever
inet6 fe80::b4d9:67ff:fe72:914c/64 scope link
valid_lft forever preferred_lft forever
Running iperf3 server:
You can then launch the iperf-client as a job file here, by default it uses the svc address (iperf3-server), but this can be overridden by setting the envvar IPERF3_HOSTNAME to the desired IP address:
oc create -f iperf-client-job.yaml
oc logs job.batch/iperf3-client
iperf3 testing: client
Running against host: 10.130.4.200
Connecting to host 10.130.4.200, port 5201
[ 5] local 10.129.3.182 port 50766 connected to 10.130.4.200 port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 109 MBytes 915 Mbits/sec 5 397 KBytes
[ 5] 1.00-2.00 sec 108 MBytes 910 Mbits/sec 0 511 KBytes
[ 5] 2.00-3.00 sec 107 MBytes 901 Mbits/sec 6 609 KBytes
[ 5] 3.00-4.00 sec 108 MBytes 910 Mbits/sec 0 730 KBytes
[ 5] 4.00-5.00 sec 108 MBytes 906 Mbits/sec 0 821 KBytes
[ 5] 5.00-6.00 sec 108 MBytes 904 Mbits/sec 0 821 KBytes
[ 5] 6.00-7.00 sec 109 MBytes 911 Mbits/sec 0 821 KBytes
[ 5] 7.00-8.00 sec 109 MBytes 911 Mbits/sec 0 821 KBytes
[ 5] 8.00-9.00 sec 108 MBytes 903 Mbits/sec 1 821 KBytes
[ 5] 9.00-10.00 sec 109 MBytes 911 Mbits/sec 0 821 KBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-10.00 sec 1.06 GBytes 908 Mbits/sec 12 sender
[ 5] 0.00-10.04 sec 1.06 GBytes 903 Mbits/sec receiver
iperf Done.
There might be occasions when you want to run more than one iperf test concurrently, this might be to test higher bandwidth NIC (A single iperf instance is limited to around 30Gbps) or to test different links in your Kube environment at the same time. Its simple to scale up the iperf server by increasing the replica count in the iperf3 server deployment descriptor. The iperf3 job yaml provided above has some rules that will avoid placing the client on the same node as the server or on which another client is running. There is a parellelism value in the job file which can be increased to run 2 or more clients in parallel. This should work (and has on occasion) as the requests to the iperf3-server service are distributed across the running servers, but sometimes errors are thrown from iperf. I suspect the request are not always distributed in an equal round-robin fashion. You could then start individual jobs each going to a seperate endpoint (oc describe svc iperf3-server).