I'm running a bunch of selenium UI tests in cent os 7 using selenium-java 2.39.0, chromedriver 2.22.
All the test cases can run smoothly in parallel(I've set parallel="tests" in the testNG xml) under the terminal command line, chrome/chromedriver can exit smoothly and successfully after all the test cases completed.
However when I run the same tets cases under jenkins 2.7.2(also tried lower version of Jenkins such as 1.644) using exactly the same test script, there are always some chrome browsers can't exit successfully after a test case completed and driver.quit was called, and gradually there will be many chrome browsers hang there. The following snapshot is an example of how the chrome browser is hanged there:
google chrome is not responding
I'm sure jenkins use the same shell, same chromedriver and same selenium when running the UI tests, just wonder how come there will be no chrome/chromedriver hang issue when running these test cases in command line, but randomly hang when run in jenkins? BTW, all the UI test cases running in jenkins that are under windows platform is also ok, it only happens in jenkins that is on a Cent OS 7 machine.