Master The Skills Of Running Chrome Headless With Selenium And Python On Linux Servers And Be Successful

Running UI Automation Tests in the cloud or on committed self facilitated Linux servers with negligible assets, will enable you to spare a few costs when you to have various tests running in parallel on a few Linux servers. It is prescribed to keep it lean and utilize Linux server disseminations without graphical UI, and to execute your experiments with a "Headless" state in Chrome. 

Headless test computerization execution mimics the genuine activities as they were preformed on a real program, yet it doesn't require any GUI. By running tests in a Headless mode you will likewise see significantly quicker execution – tests that keep running in a headless mode run 2 to 10 times speedier than running in a typical mode program. The reason is basic – In a Headless mode the time it takes to stack the JavaScript, CSS and render the HTML are much lower since it isn't beginning up a program GUI. 

What's more, you should need to keep creating while at the same time running a 10 minute UI test suite content on your machine, without having the program fly up at regular intervals. With a content running in a Headless mode you can do only that – Run the content from the order line and keep creating, rather than being stuck watching your screen for the following 10 minutes. 

For this article I will utilize CentOS as my Linux server, despite the fact that a similar objective can be accomplished on most Linux situations. 

Initially, you will require Python and Selenium on your Linux machine: 

pip is the bundle administration framework for Python. 

yum introduce python27 

yum introduce python-pip 

pip introduce - U selenium 

All in all, how might you run UI Automation Tests in a Headless mode? 

These are the 2 approaches to handle this issue (If you need to utilize Firefox webdriver, you just need to utilize the main alternative): 

PyVirtualDisplay. 

Chrome headless banner (in chrome_options). 

Note: If your site has SSL blunders, overlooking authentication mistakes does not work exceptionally well with chromedriver. I would propose utilizing Firefox. This line of code won't help: 

chrome_options.add_argument('- - overlook declaration mistakes') 

PyVirtualDisplay is a Python wrapper for Xvfb, Xephyr and Xvnc. You should have Xvfb introduced on your Linux machine and additionally the PyVirtualDisplay library: 

yum introduce Xvfb 

pip introduce PyVirtualDisplay 

Xvfb (X virtual framebuffer) is a show server executing the X11 show server convention. Xvfb executes every graphical task utilizing the virtual memory without demonstrating any screen yield, not at all like other show servers. Also, Xvfb does not require the machine to have a screen, illustrations connector or information gadget. The main prerequisite is a system layer. 

Default show ought to be :1 

You can check this with the summon: 

env | grep DISPLAY 

You should ensure your intermediary settings (if required) are designed effectively, since off base arranged intermediaries can make ChromeDriver not work legitimately. Since ChromeDriver takes the intermediary settings from the env variable, you can set or disconnected the intermediaries as you wish (Remove them in case you're running tests on nearby machines): 

unset http_proxy 

unset https_proxy 

Or on the other hand: 

send out http_proxy=10.0.x.x 

send out https_proxy=10.0.x.x 

You could likewise include the intermediary settings by means of chrome_options or desired_capabilities for Firefox: 

# Firefox: 

myProxy = "10.0.x.x:yyyy" 

intermediary = Proxy({ 

'proxyType': ProxyType.MANUAL, 

'httpProxy': myProxy, 

'ftpProxy': myProxy, 

'sslProxy': myProxy, 

'noProxy': '' 

}) 

driver = webdriver.Firefox(proxy=proxy) 

# Chrome: 

myProxy = "10.0.x.x:yyyy" 

chrome_options = webdriver.ChromeOptions() 

chrome_options.add_argument('- - intermediary server=%s' % myProxy ) 

Refresh chrome to its most recent adaptation to be perfect with the most recent ChromeDriver twofold. 

Utilize this summon: 

yum refresh google-chrome 

Download ChromeDriver and duplicate to your envelope of decision: 

cd/home/dev 

wget https://chromedriver.storage.googleapis.com/2.35/chromedriver_linux64.zip 

unfasten chromedriver_linux64.zip 

Furthermore, run your content of decision: 

from selenium import webdriver 

# Option 1 - with ChromeOptions 

chrome_options = webdriver.ChromeOptions() 

chrome_options.add_argument('- - headless') 

chrome_options.add_argument('- - no-sandbox') # required when running as root client. else you would get no sandbox mistakes. 

driver = webdriver.Chrome(driver_path='/home/dev/chromedriver', chrome_options=chrome_options, 

service_args=['- - verbose', '- - log-path=/tmp/chromedriver.log']) 

# Option 2 - with pyvirtualdisplay 

from pyvirtualdisplay import Display 

show = Display(visible=0, size=(1024, 768)) 

display.start() 

driver = webdriver.Chrome(driver_path='/home/dev/chromedriver', 

service_args=['- - verbose', '- - log-path=/tmp/chromedriver.log']) 

# Log way added by means of service_args to see blunders if something turns out badly (dependably a smart thought - huge numbers of the mistakes I experienced were depicted in the logs) 

# And now you can include your site/application testing usefulness: 

driver.get('https://python.org') 

print(driver.title) 

# driver.click... 

Add the code to a test.py record, and run it: 

python test.py