[🐛 Bug]: Xvfb and VNC crash under M1 Mac when using AMD64 image #2298
Comments
|
@slhck, thank you for creating this issue. We will troubleshoot it as soon as we can. Info for maintainersTriage this issue by using labels.
If information is missing, add a helpful comment and then
If the issue is a question, add the
If the issue is valid but there is no time to troubleshoot it, consider adding the
If the issue requires changes or fixes from an external project (e.g., ChromeDriver, GeckoDriver, MSEdgeDriver, W3C),
add the applicable
After troubleshooting the issue, please add the Thank you! |
|
This issue has been reported several times. Please check the comments of the reports history and if you are still facing it after trying out the different suggestions, report it to the Chrome folks. |
diemol
added
G-chromedriver
|
Hi, @slhck. You can see if the feature is passing in the Web Platform Tests. If it is something new, please create an issue with the ChromeDriver team. |
|
Thanks for the quick answer, and sorry, I somehow didn't see those. Based on my Internet searches most issues had to do with the Can you please let me know which specific solution from these other issues you had in mind? Because you are of course right about the same error message, and quick to close the issue, yet the problem is only occurring after a first successful test. It's not like this issue is exhibited from the first startup. Specifically, after the first session is stopped, I see: If indeed XVFB is not running, I would expect Chrome not being able to even launch. So this seems to be a different issue than the other ones — just the same final error message. More specifically:
|
|
This one is also giving information about this SeleniumHQ/selenium#14076 |
|
What specifically do you mean with respect to the other issue? As far as I can tell the session itself is closed properly in my case, and the resource consumption is normal. It's also just one test, so it's not only occurring with multiple parallel executions like the OP mentioned. The underlying reason seems to be the VNC and XVFB session crashing. What could cause that? I have also tried with more shared memory (4G) but that did not help. |
|
What do you see in the XVFB and VNC logs? https://github.com/SeleniumHQ/docker-selenium/blob/trunk/NodeBase/selenium.conf#L13 |
|
When I dump the Xvfb logs during the first test I get: The "explicit kill or server shutdown" message appears when the test has finished. No subsequent sessions can be started. I wonder if the supervisord configuration should be changed to restart Xvfb in case of failure? Still, I do not see any explicit calls to |
|
Are you using a macOS M1/M2/M3? |
|
Yes, I am using a M1 Mac. This is from VNC: Apparently it's the: So apparently running the Linux/AMD64 image itself with Chrome is not possible? When I run the Chromium image natively, it works: |
|
You are using an x86 image on ARM |
slhck
changed the title
|
Yes, although there is no information that this wouldn't work, as x64 is just being emulated here, and so far I have also been able to run other images if they weren't available as multi-arch images. I think that the README is a bit unclear in that regard. I fixed the title of the issue and will provide a PR to update the README. |
|
This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs. |
What happened?
I am running the standalone Chrome Docker image to run some tests against. The first test always works fine, but any subsequent tests fail with an error:
My test consists of a simple set of steps:
When I run the test once, all is fine. The second time I run it, it errors out.
Command used to start Selenium Grid with Docker (or Kubernetes)
docker run --platform linux/amd64 -p 4444:4444 -p 7900:7900 --shm-size="2g" selenium/standalone-chrome:126.0-20240621Relevant log output
Operating System
macOS
Docker Selenium version (image tag)
126.0-20240621
Selenium Grid chart version (chart version)
No response
The text was updated successfully, but these errors were encountered: