Locust on OpenShift

Run distributed load tests using Locust and OpenShift.

Original idea and inspiration came from: https://medium.com/locust-io-experiments/locust-io-experiments-running-in-kubernetes-95447571a550

Seed the project with your data to test

Run the following command to move the test data to the cluster.

$ ./seed.sh file_name_with_test.py https://www.url-to-test.com

Note: the seed.sh use the oc command to push the changes.

Install OC command

• Download oc Client Tools

$ wget https://github.com/openshift/origin/releases/download/v3.11.0/openshift-origin-client-tools-v3.11.0-0cbc58b-linux-64bit.tar.gz

• Extract the downloaded .tar.gz

$ tar -xf openshift-origin-client-tools-v3.11.0-0cbc58b-linux-64bit.tar.gz

• Move oc binary file to a directory in your local machine path

cd openshift-origin-client-tools-v3.11.0-0cbc58b-linux-64bit
$ mv oc /usr/local/bin/

• Apply Executable Permissions to the Binary

$ chmod +x /usr/local/bin/oc

Install on OpenShift

• Login to the OpenShift Cluster

$ oc login https://www.cluster-address.com

• Create new project:

$ oc new-project locust --display-name=Locust --description="Locust.io load tests cluster";

• Select project namespace:

$ oc project locust;

• Deploy Locust master pod:

$ oc process -f master-deployment.yaml | oc create -f -

• Deploy Locust slave pod:

$ oc process -f slave-deployment.yaml | oc create -f -

Note: the slave-deployment.yaml has the comment out code for setting up auto scaling the pods, but Locust will reset the test to distribute the load every time a new slave is added. You can manually create the slave pods before starting the test and destroy them when done.