Facing "Net::ReadTimeout" Issue in GitHub Actions CI/CD Environment

[nishanthmathiyazhagan9704]

Select Topic Area

Show & Tell

Body

Hi GitHub Community,

I hope you're all doing well. I'm currently facing an issue with "Net::ReadTimeout" while running my tests in GitHub Actions CI/CD environment. I've tried various troubleshooting steps, but the problem persists, and I'm seeking some assistance from the community.

Issue Description:

1. The error occurs during test execution in the GitHub Actions CI/CD environment.
2. The tests are written in Ruby using Capybara and Selenium WebDriver.
3. The error message specifically mentions "Net::ReadTimeout" with "#Socket:(closed)".
4. I'm using Capybara, Selenium, Firefox, and Geckodriver

What I've Tried So Far:

1. Increased the default wait time for Capybara using Capybara.default_max_wait_time.
2. Checked the URLs and verified that they are accessible.
3. Ensured that JavaScript is executing without errors on the pages.
4. Tested the same scenario in non-headless mode to observe the execution in local
5. Verified network connectivity and GitHub Actions environment settings.

However, despite these efforts, the issue persists, and I'm unable to pinpoint the root cause of the "Net::ReadTimeout" error.

I'm looking for any insights, suggestions, or guidance from the community on how to resolve this issue. If you've encountered a similar problem before or have any ideas on how to troubleshoot and resolve "Net::ReadTimeout" in GitHub Actions, I would greatly appreciate your input.

Here's where I am facing the issue
Capybara.current_session.driver.browser.manage.window.maximize

Failures:

  1) Crawl all pages of the website takes a snapshot of all pages
     Failure/Error: Capybara.current_session.driver.browser.manage.window.maximize

     Net::ReadTimeout:
       Net::ReadTimeout with #<Socket:(closed)>

Environment Details:

1. Operating System: macOS Monterey 12.1
2. Browser: 115.0.2
3. GitHub Actions Environment: ubuntu-latest

Thank you in advance

[devrose04]

Thank you

[faisalnazir7]

Instead of just increasing the default wait time (Capybara.default_max_wait_time), you can try explicitly setting a wait time for the specific action that encounters the timeout issue. For example, you can use Capybara.using_wait_time to set a longer wait time for the maximize action:

Capybara.using_wait_time(30) do
  Capybara.current_session.driver.browser.manage.window.maximize
end

[nishanthmathiyazhagan9704]

Hi @faisalnazir7 Thank you for the response. I appreciate your help in trying the explicit wait suggestion. However, it seems that the timeout issue persists despite these efforts. I'm still seeking a resolution to this problem.

Failures:

  1) Crawl all pages of the website takes a snapshot of all pages
     Failure/Error: Capybara.current_session.driver.browser.manage.window.maximize

     Net::ReadTimeout:
       Net::ReadTimeout with #<Socket:(closed)>
     # ./spec/crawler_spec.rb:12:in `block (2 levels) in <top (required)>'

Finished in 1 minute 1.76 seconds (files took 0.67244 seconds to load)
1 example, 1 failure

[robacarp]

@nishanthmathiyazhagan9704 were you ever able to resolve your issue here?

[sebgrebe]

@nishanthmathiyazhagan9704 or @robacarp did one of you solve this? We've been facing this issue suddenly since last week, with a similar setup to yours:

• Github Actions (ubuntu)
• Capybara
• Selenium
• Chrome

[robacarp]

Sorry, no. We boiled it down to a possible race condition in Capybara, which is somewhat documented here. The capybara team seems to be ignoring it, and someone along the way suggested trying Playwright instead of Selenium, which may be our next attempt. At any rate, our problem isn't an issue with Github Actions.

[sebgrebe]

Thanks for the pointer!