Update 2021-04-20: this affects all ....OnOutOfMemoryError switches, not only CrashOnOutOfMemoryError.

--------------


FULL PRODUCT VERSION :
java version "1.8.0_92"
Java(TM) SE Runtime Environment (build 1.8.0_92-b14)
Java HotSpot(TM) 64-Bit Server VM (build 25.92-b14, mixed mode)

FULL OS VERSION :
Linux xxxxx 4.0.9 #1 SMP Fri Feb 26 15:21:14 PST 2016 x86_64 x86_64 x86_64 GNU/Linux
Darwin xxxxx 15.4.0 Darwin Kernel Version 15.4.0: Fri Feb 26 22:08:05 PST 2016; root:xnu-3248.40.184~3/REL

A DESCRIPTION OF THE PROBLEM :
The newly introduced CrashOnOutOfMemoryError option doesn't appear to work when the cause of the OOM is due to not being able to create new threads.

THE PROBLEM WAS REPRODUCIBLE WITH -Xint FLAG: Did not try

THE PROBLEM WAS REPRODUCIBLE WITH -server FLAG: Did not try

STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
Run the program below with "java -XX:+CrashOnOutOfMemoryError Test". After a few seconds, an OutOfMemoryError (unable to create new native thread) will be printed to the console but the program won't exit.

EXPECTED VERSUS ACTUAL BEHAVIOR :
I would expect the VM to crash, but it keeps running.
REPRODUCIBILITY :
This bug can be reproduced always.

---------- BEGIN SOURCE ----------
class Test {
    public static void main(String[] args)
            throws InterruptedException
    {
        Thread t = new Thread(() -> {
            while (true) {
                new Thread(() -> {
                    try {
                        Thread.currentThread().join();
                    }
                    catch (InterruptedException e) {
                    }
                }).start();
            }
        });

        t.start();
        Thread.currentThread().join();
    }
}
---------- END SOURCE ----------

CUSTOMER SUBMITTED WORKAROUND :
Use a JVMTI agent to intercept ResourceExhausted events and forcibly exit the VM (e.g., https://github.com/airlift/jvmkill)