I started out trying to learn how to write automated tests for a small project but nothing was working right out of the box. After a couple hours of searching & experimenting I found the right configuration for my project & figured sharing it might help folks in the future.
Here's a small summary of the errors I encountered on this debugging journey:
Using Jasmine & WDIO, send_keys was crashing
It was a Firefox/geckodriver bug, or something like that
WDIO appeared to hang after switching from Firefox to Chrome
Chrome needed to be run in --no-sandbox mode, essentially
I figured my problem was probably stemming from having WDIO execute my tests on my local machine while Selenium was hammering on the browser in a Vagrant VM. So this will mainly be applicable for people using separate environments (vagrant->local, vagrant->vagrant, docker->local, etc) for WDIO & Selenium/Chromedriver. Here is a gist of the configuration file I ended up with.
I started with fanatique/vagrant-selenium-vm and modified it to use Chrome instead of Firefox because send_keys was broken with Firefox/geckodriver at the time of writing. After swapping out the packages, I'd start the tests with wdio but it would appear to hang. Turning on verbose logging showed it trying to start but failing with no explanation why.
[00:06:39] COMMAND POST "/wd/hub/session"
[00:06:39] DATA {"desiredCapabilities":{"javascriptEnabled":true,"locationContextEnabled":true,"handlesAlerts":true,"rotatable":true,"browserName":"chrome","loggingPrefs":{"browser":"ALL","driver":"ALL"},"requestOrigins":{"url":"http://webdriver.io","version":"4.6.2","name":"webdriverio"}}}
It took removing the & from the nohup java ... command in fanatique/vagrant-selenium-vm's setup.sh to see the logs from Selenium in real time, then I was able to see a "only local connections are allowed" message from chromedriver. That led me to a SO post that said to add --whitelisted-ips="" as an arg to chromedriver - but I was still getting the local connections error message. Chrome itself ended up needing a --no-sandbox flag - that allowed WDIO to connect to the chromedriver in Selenium and my tests ran from there.
Related
I am working on an integration test, using Selenium/Java/Cucumber. When I run the test on localhost, or remotely on a VM, it passes fine. But when it runs as part of a Jenkins pipeline it hangs with the error "Timed out receiving message from renderer". I have googled this and there are several possible causes, including version conflicts. But I can't seem to figure out how to check these versions in the jenkins pipeline. Any pointers on what to look into would be helpful. I believe the tests are running headless on the pipeline, because there is no Chromedriver pluigin installed in our Jenkins. The test is pretty straightforward - it involves clicking on an element within an iframe (this is where it hangs) and checking for some output. I have even tested it running manually on the VM created by the very same pipeline and it always works without an issue.
I have a strange problem, which I'm unable to solve for sometime. I have a selenium WDS for Firefox browser. I tried to execute the code in JMeter UI, everything works fine. But when I try to execute the same in a server (using XVFB driver), I get the error:
rm="sun.org.mozilla.javascript.EvaluatorException: Can't find method org.openqa.selenium.remote.RemoteWebDriver.executeScript(string,org.openqa.selenium.remote.RemoteWebElement). (<Unknown source>#17) in <Unknown source> at line number 17"
All the required drivers are the same as the one I locally try. Yet, I get the error only when I run in the server but not locally. Please help
Most probably you have different versions of either WebDriver Sampler or dependency libraries, I don't see any other ways for the same script to work on one machine and don't work on another.
Install the same version of JMeter onto both machines (use JMeter 3.2 or later, whatever is available at JMeter Downloads Page)
Install latest version of the WebDriver Sampler plugin using JMeter Plugins Manager
Your test should now be running fine.
I have a project suite of automation tests that uses Protractor and Jasmine. As of a couple months ago, my tests stopped working in FireFox. Everything works perfectly in Chrome, and almost perfectly in IE.. because why would IE work ya know.
There have been multiple posts about this and I've read through all of them, but I haven't been able to determine if there's a workaround/solution to this. It's not entirely a HUGE deal if my tests run in FireFox, but with a web application it's mighty nice to be able to test all instances.
Basically what happens when I run a test in FireFox:
Command prompt acts like everything is about to run properly (FireFox driver starts up.. server remains running etc etc)
Firefox opens for a split second (icon appears on the task bar, browser flashes on the screen and closes)
After a good minute or so protractor gives me the following message:
[firefox #01-0] WebDriverError: Unable to connect to host XXX on port XXX
after 45000 ms.
With a ton of DEBUG lines referring to JSON files and Firefox addons.
Any thoughts or other sources I could check out would be sweet. Many thanks in advance.
Problem is with latest Firefox support from latest selenium-server. Use geckoDriver and it will work.
Follow this URL
Trying to port selenese commands from the Selenium IDE to a cron-usable task, and have gotten as far as getting these to run with Webdriver's HTMLSuite command using a local installation of Firefox. However, we need these to run without opening a visual browser - as we're already using PhantomJS elsewhere in the application this seemed like the logical option rather than going down the XVFB route.
Is this possible? I've tried a variety of commands but keep getting errors of phantomjs not being an accepted browser.
I have a problem with Selenium under Jenkins 1.446 together with Xvfb: It looks like firefox is started correctly because I let my failing tests record screenshots. These screenshots all show the same failing page, which in my case is the starting page where the tests should begin. So I gues the selenium WebDriver commands do not arrive. What could be the reason? By the way, the tests are running perfectly on my local machine.
I'm using Firefox 9.0.1 with no specific test profile and no AddOns, Ubuntu 10.04, Senlenium 2.16.1
Log entry: com.thoughtworks.selenium.SeleniumException: Timed out waiting for action to finish
Thanks!
EDIT: Issue seems to be fixed by a system reboot ...
I experienced a very similar issue with FF9.0.1 and Selenium 2.16.1 running through JUnit launched by Maven SureFire plugin run by Jenkins on a WinXP node.
2.17.0 fixed the issue for me. Try updating to the latest Selenium.
Through RDP, I was able to watch the tests running. The tests appeared to be partially blocked by a prompt from Firefox asking about collecting anonymous usage statistics. Manually answering the prompt would allow the test to continue but because Selenium creates a new profile each time by default, the prompt would return on the next browser launch. Running the tests locally with a pre-configured FF profile allowed me to persist that the prompt had been answered. This isn't possible on my XP node because the tests are running as 'System'.
Selenium/WebDriver would normally take care of this for you by marking the prompt as already answered in a temporary FF profile configuration but a bug was causing the value to be set to the wrong value. You can inject a profile configuration to the Selenium FirefoxDriver driver to pragmatically configure stuff like this prompt but the bug appeared to prevent this as well. This has been resolved in Selenium 2.17.0 (http://selenium.googlecode.com/svn/trunk/java/CHANGELOG see 2.17.0 WebDriver bug fixes).