@shubha-rajan if you have time can you take a look at and help with this?

@shubha-rajan if you have time can you take a look at and help with this?

Sure! I saw that you tried changing the fork count in a previous commit. I think you were on the right track but also need to add this line: <argLine>-Xmx1024m -XX:MaxPermSize=256m</argLine>

Source:
https://stackoverflow.com/questions/23260057/the-forked-vm-terminated-without-saying-properly-goodbye-vm-crash-or-system-exi

Sidenote: This is a bit concerning https://maven.apache.org/surefire/maven-surefire-plugin/faq.html#vm-termination

If this happens on a CI environment and only after it runs for some time, there is a fair chance your test suite is leaking some kind of OS-level resource that makes things worse at every run.

For Java 17, the image needs to be updated. I've added a PR googleapis/testing-infra-docker#206. As for the potential leak, I unfortunately would lean towards waiting to see if we move to GitHub Actions before debugging Kokoro.

The forks seem to be causing issues. We may need to skip the hello-broken tests for J17 since I need to set env vars for the tests and this is causing the main issue.