On Thu, 17 Jul 2025 08:04:50 GMT, Richard Reingruber <rr...@openjdk.org> wrote:
>> test/hotspot/jtreg/serviceability/HeapDump/UnmountedVThreadNativeMethodAtTop.java
>> line 59:
>>
>>> 57: @BeforeEach
>>> 58: void doGC() {
>>> 59: System.gc();
>>
>> The Systen.gc() might not trigger full GC.
>>
>> While not tjust o restrict heap for this test? 256M should be enough.
>
>> The Systen.gc() might not trigger full GC.
>
> Indeed. Might be better to use the Whitebox or start a dedicated VM.
>
>> While not tjust o restrict heap for this test? 256M should be enough.
>
> This would not help. The tests accumulate heap dumps until OOM. See
> reproducer.
It might not be the perfect solution but it surely reduces the noise in our
testing.
So let's have this improvement in jdk25.
In jdk26 we can change the test to use
[WhiteBox::fullGC](https://github.com/openjdk/jdk/blob/18190519e73705281adf3f94d710d000e75b1729/test/lib/jdk/test/whitebox/WhiteBox.java#L561).
What do you think?
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/26343#discussion_r2212657017
