RFR: 8192983: gc/g1/TestVerifyGCType.java might fail on loaded machines

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

RFR: 8192983: gc/g1/TestVerifyGCType.java might fail on loaded machines

Stefan Johansson
Hi,

Please review this test fix for:
https://bugs.openjdk.java.net/browse/JDK-8192983

Webrev:
http://cr.openjdk.java.net/~sjohanss/8192983/00/

Summary:
On big or loaded machines the tests was to fragile to always pass. Have
added restrictions on the heap size and also made sure the test complete
the concurrent cycle before the young GCs are triggered.

Testing:
* mach5 with 10 times repeat without any problems
* mach5 hotspot_gc test list with -Xcomp (failed in PIT) currently running

Thanks,
Stefan
Reply | Threaded
Open this post in threaded view
|

Re: RFR: 8192983: gc/g1/TestVerifyGCType.java might fail on loaded machines

Thomas Schatzl
Hi,

On Tue, 2017-12-05 at 11:47 +0100, Stefan Johansson wrote:

> Hi,
>
> Please review this test fix for:
> https://bugs.openjdk.java.net/browse/JDK-8192983
>
> Webrev:
> http://cr.openjdk.java.net/~sjohanss/8192983/00/
>
> Summary:
> On big or loaded machines the tests was to fragile to always pass.
> Have
> added restrictions on the heap size and also made sure the test
> complete
> the concurrent cycle before the young GCs are triggered.
>
> Testing:
> * mach5 with 10 times repeat without any problems
> * mach5 hotspot_gc test list with -Xcomp (failed in PIT) currently
> running

line 251: comment references not-used flag
-XX:+ExplicitGCIgnoresConcurrent

Otherwise looks good.

Thanks,
  Thomas
Reply | Threaded
Open this post in threaded view
|

Re: RFR: 8192983: gc/g1/TestVerifyGCType.java might fail on loaded machines

Stefan Johansson
Hi Thomas,

On 2017-12-05 11:52, Thomas Schatzl wrote:

> Hi,
>
> On Tue, 2017-12-05 at 11:47 +0100, Stefan Johansson wrote:
>> Hi,
>>
>> Please review this test fix for:
>> https://bugs.openjdk.java.net/browse/JDK-8192983
>>
>> Webrev:
>> http://cr.openjdk.java.net/~sjohanss/8192983/00/
>>
>> Summary:
>> On big or loaded machines the tests was to fragile to always pass.
>> Have
>> added restrictions on the heap size and also made sure the test
>> complete
>> the concurrent cycle before the young GCs are triggered.
>>
>> Testing:
>> * mach5 with 10 times repeat without any problems
>> * mach5 hotspot_gc test list with -Xcomp (failed in PIT) currently
>> running
> line 251: comment references not-used flag
> -XX:+ExplicitGCIgnoresConcurrent
Will clean that out before pushing.

Thanks,
Stefan

> Otherwise looks good.
>
> Thanks,
>    Thomas

Reply | Threaded
Open this post in threaded view
|

Re: RFR: 8192983: gc/g1/TestVerifyGCType.java might fail on loaded machines

Stefan Johansson
Got an offline review from Jesper and pushed this fix since it was an
integration blocker.

Thanks,
Stefan

On 2017-12-05 11:55, Stefan Johansson wrote:

> Hi Thomas,
>
> On 2017-12-05 11:52, Thomas Schatzl wrote:
>> Hi,
>>
>> On Tue, 2017-12-05 at 11:47 +0100, Stefan Johansson wrote:
>>> Hi,
>>>
>>> Please review this test fix for:
>>> https://bugs.openjdk.java.net/browse/JDK-8192983
>>>
>>> Webrev:
>>> http://cr.openjdk.java.net/~sjohanss/8192983/00/
>>>
>>> Summary:
>>> On big or loaded machines the tests was to fragile to always pass.
>>> Have
>>> added restrictions on the heap size and also made sure the test
>>> complete
>>> the concurrent cycle before the young GCs are triggered.
>>>
>>> Testing:
>>> * mach5 with 10 times repeat without any problems
>>> * mach5 hotspot_gc test list with -Xcomp (failed in PIT) currently
>>> running
>> line 251: comment references not-used flag
>> -XX:+ExplicitGCIgnoresConcurrent
> Will clean that out before pushing.
>
> Thanks,
> Stefan
>
>> Otherwise looks good.
>>
>> Thanks,
>>    Thomas
>