DevOps
Running Selenium tests from the command line on a Delivery Pipeline
August 11, 2017 | Written by: Paul Webster
Share this post:
When creating a cloud application, you often want to have automated tests for your basic UI functionality, like Selenium based tests. You can run Selenium tests through Sauce Labs or you can run them locally.
However, when you run them in a Delivery Pipeline in a DevOps toolchain on IBM Bluemix, you might see errors because the server doesn’t have a way to display the browser window.
This example creates a toolchain with a pipeline job that installs a virtual display buffer that allows Firefox to display. Then the simple python example will demonstrate running Selenium tests in the Delivery Pipeline.
Running tests on a Delivery Pipeline
To run Selenium tests in Delivery Pipeline, there are 4 main steps:
- Install the browser and virtual display buffer.
- Install the Selenium drivers for your test language.
- Run the virtual display buffer.
- Run the tests.
Install the browser and virtual display buffer
Make sure your build environment includes both tools that are needed. In this example, I’m using Firefox.
- Install Firefox
sudo apt-get -y --force-yes update sudo apt-get -y --force-yes install firefox firefox -v
- Install Xfvb and some supporting fonts
sudo apt-get -y --force-yes install xvfb xfonts-100dpi \ xfonts-75dpi xfonts-cyrillic xorg dbus-x11
Install the Selenium drivers for your test language
In this example, the python Selenium package uses the geckodriver to drive the Firefox browser. Both the python Selenium package and geckodriver need to be installed.
- Install the geckodriver
wget https://github.com/mozilla/geckodriver/releases/download/v0.17.0/geckodriver-v0.17.0-linux64.tar.gz mkdir bin tar xvzf geckodriver-v0.17.0-linux64.tar.gz -C bin export PATH=$(pwd)/bin:$PATH
- Install the python package
sudo pip install -U Selenium
Run the virtual display buffer
The virtual display buffer program allows Firefox to behave like it’s displaying on your screen. The :10 indicates
what screen number the virtual display runs on
Xvfb :10 -ac & export DISPLAY=:10
Run the tests
Now you can run the tests using whatever Selenium bindings you have installed.
py.test --junitxml results.xml tests/test_python_search.py || :
Try it
You can check out an example of a toolchain that builds a python test file and runs the test by clicking
the Create Toolchain button.
Troubleshooting
Scenario: None of the Selenium tests run, and you see either of these error messages:
WebDriverException: Message: The browser appears to have exited before we could connect. If you specified a log_file in the FirefoxBinary constructor, check it for details.
WebDriverException: Message: Can't load the profile. Profile Dir: %s If you specified a log_file in the FirefoxBinary constructor, check it for details.
Explanation: Your versions of Firefox and the Python Selenium driver might not be compatible. This problem is common.
Possible solution: Load either an older version of Firefox or the newest version.
- To load an older version, enter this command:
sudo apt-get install firefox=28.0+build2-0ubuntu2
- To use the newest version of Firefox, enter this command:
sudo add-apt-repository -y ppa:mozillateam/firefox-stable sudo apt-get -y update
If you need to troubleshoot other issues, install a virtual machine (VM) by using a tool such as Vagrant. The Delivery Pipeline runs on Ubuntu 14.04, so a VM with that version of Ubuntu is suitable for testing and can save you time, as you won’t need to run a build every time that you want to test.
Two Tutorials: Plan, Create, and Update Deployment Environments with Terraform
Multiple environments are pretty common in a project when building a solution. They support the different phases of the development cycle and the slight differences between the environments, like capacity, networking, credentials, and log verbosity. These two tutorials will show you how to manage the environments with Terraform.
Transforming Customer Experiences with AI Services (Part 1)
This is an experience from a recent customer engagement on transcribing customer conversations using IBM Watson AI services.
Analyze Logs and Monitor the Health of a Kubernetes Application with LogDNA and Sysdig
This post is an excerpt from a tutorial that shows how the IBM Log Analysis with LogDNA service can be used to configure and access logs of a Kubernetes application that is deployed on IBM Cloud.