Unknow reason for Segmentation fault (core dumped)

[peterpengnz]

Hi
I am trying to use Ferrum in my Rails project, which is hosted as ECS in AWS.
When using the function, I got this error message in rails log:

Ferrum::ProcessTimeoutError Browser did not produce websocket url within 10 seconds, try to increase `:process_timeout`. See https://github.com/rubycdp/ferrum#customization

My code looks like this:

browser = Ferrum::Browser.new(
        browser_options: { "no-sandbox": nil },
        window_size: [viewport_width, viewport_height],
        js_errors: true,
        timeout: 30
      )

I tried adding process_timeout with a much longer value. However, the issue remains.
There is a core.xxx (xxx is the process id) was dumped in the rails project directory.
I used the top command while the Chromium process started, and I can see it is running by using the same user as the Rails project.

If I try to run this command, it works:

ip-x-x-0-31:~$ chromium --no-sandbox --version
Chromium 136.0.7103.113 Alpine Linux

And if as root user:

ip-x-x-0-31:/app# chromium --no-sandbox  --version
Segmentation fault (core dumped)

The container has enough free memory (2.5G free) and it is using Alpine image:

# Fresh alpine image
FROM ruby:3.4.4-alpine3.21 as base

Any hit why this is happening and how to debug?

Thanks

[route]

You have no memory left

[peterpengnz]

thanks!