Participants [Chrome] in headless mode
tusharsonawanes opened this issue · 2 comments
Hello,
I am following this tutorial for setting up my testing environment. The Selenium grid machine is where I have my jdk and the selenium grid configured. All the other nodes ( Ubuntu 18.04, GUI, chrome) have joined the same successfully. ( Not using the root user )
When I configure 30 chrome instances in every node, I can successfully see them listed in my selenium grid endpoint.
Problem.
When I run the tests from the selenium node using this command
./scripts/malleus.sh --conferences=1 --participants=90 --senders=1 --audio-senders=2 --duration=1000 --room-name-prefix=hamertesting --hub-url=http://seleniumhub:4444/wd/hub --instance-url=https://myjitsiinstallation
I can see that only ~50-60 chrome instances join the meeting. Moreover, I see that the 50-60 pariticipants ( Chrome instances ) get opened on my Laptop. When I keep the task manager open, I see that the CPU utilization is full at 60th participant because of which java throws an exception and the further participants ( chrome instances ) cannot start.
What is expected
Can I have the 50-60 chrome instances running in a headless mode ? I can just go to the instance url and see all the participants listed there. ( The 50-60 chrome instances should not open on my laptop )
The chrome instances should be running on the nodes, not locally on the machine where you run torture
jitsi-meet-torture/scripts/malleus.sh
Line 159 in c1fcdcd
That is strange
Exactly. I see a lot of chrome instances in my laptop.
Just a question :
Is it happening because I take VM1 shell from my laptop on which I have selenium grid + the torture code running ?