0

I'm using a Centos 7 server to run automatic e2e test by Jenkins. My stack consist in protractor test using webdriver-manager. Because i want to automate the process, i create a service to run webdriver from start. But after few builds failed, the webdriver increase the use of memory over my limit (2 GB physic + 2 GB virtual) and the server starts to fail. Any ideas or workaround?

I'm using webdriver-manager 10.1.0 with chromedriver 2.22, Selenium-standalone 2.53, and jenkins 2.10

cnishina
  • 5,016
  • 1
  • 23
  • 40
Gustavo Saavedra
  • 1
  • My impĺementention use xvfb to run headless and the webdriver is used by few different task to run at the same time (in this moment, 2) – Gustavo Saavedra Jul 05 '16 at 21:48
  • 1
    I have similar issue related with spec of the server which executing my e2e test. Even running on headless browser, it is still a heavy task, you need a decent spec to run it properly. For my case, 4 processors CPU (2.2GHz each), 4GB RAM, 128MB VRAM. – Linh Pham Jul 06 '16 at 04:04
  • Are you using the local driverprovider feature? Or are you using `webdriver-manager start` and setting your Protractor config to `seleniumAddress: 'http://127.0.0.1:4444/wd/hub'`? Also, if you do `ps aux | grep selenium` how many instances of chromedriver, selenium-standalone are you seeing? – cnishina Dec 02 '16 at 09:53
  • Please try the most up to date version. There was a bug that left chromedriver sessions opened. Thanks! – cnishina Sep 11 '18 at 09:27

0 Answers0