RFR 8230055: ModuleStressGC.java times out on Win*
Daniel D. Daugherty
daniel.daugherty at oracle.com
Fri Nov 8 14:29:45 UTC 2019
Chris Plummer has been chasing some test failures where some system
calls just take a really long time to complete. I _think_ he has
isolated his sightings to a specific system calls and machine configs
in our test farm.
You might want to check with him to see if your test is running into
something similar...
Dan
On 11/8/19 7:59 AM, Harold Seigel wrote:
> Thanks Leonid.
>
> I'll withdraw the webrev.
>
> Harold
>
> On 11/7/2019 11:32 PM, Leonid Mesnik wrote:
>> Hi
>>
>> I am not sure that it is a good idea. The bug says that test times
>> out and take more 40 minutes.
>>
>> However usually it takes less than 2 minutes to complete test (on any
>> platform). So I think that it is not just a long test but rather some
>> outlier which should be investigated.
>>
>> Leonid
>>
>> On 11/7/19 2:03 PM, Harold Seigel wrote:
>>> Hi,
>>>
>>> Please review this small change to help prevent test
>>> runtime/modules/ModuleStress/ModuleStressGC.java from timing out.
>>> The change reduces the number of loop iterations in the test by 40%.
>>>
>>> Open Webrev:
>>> http://cr.openjdk.java.net/~hseigel/bug_8230055/webrev/index.html
>>>
>>> JBS Bug: https://bugs.openjdk.java.net/browse/JDK-8230055
>>>
>>> The change was tested by running Mach5 tier2 tests on Linux-x64,
>>> Solaris, Windows, and Mac OS X.
>>>
>>> Thanks, Harold
>>>
More information about the hotspot-runtime-dev
mailing list