RFR(xs): 8161552: Test issue: VM init failed: GC triggered before VM initialization completed. Try increasing NewSize, current value 768K.
dmitry.fazunenko at oracle.com
Thu Jul 21 12:05:23 UTC 2016
On 21.07.2016 14:17, Thomas Schatzl wrote:
> On Thu, 2016-07-21 at 13:34 +0300, Dmitry Fazunenko wrote:
>> On 21.07.2016 12:30, Thomas Schatzl wrote:
>>> On Wed, 2016-07-20 at 20:01 +0300, Dmitry Fazunenko wrote:
>>>> I'm looking for a couple of reviewers for a simple fix of two
>>>> The tests check some basic assertions when VM is started with max
>>>> size set to some small values.
>>>> The tests work almost everywhere, but there are cases when with
>>>> GC triggered before VM initialization completed.
>>>> 2M is valid but extremely small max heap size. Cases of usages
>>>> small heap are rather theoretical than practical.
>>>> To fix the problem I made the tests weaker, the cases setting max
>>>> to 2M are removed.
>>>> Testing: locally + RBT
>>> looks good, but I think this decision should be more prominently
>>> stated in the test.
>>> Particularly the comments in lines 45-47 need to be updated, and
>>> 49 to 53 also refer to now apparently outdated behavior.
>> Thanks for catching this, I missed that comments.
>> Updated version:
> - line 37 contradicts with the new line 54 now. It's probably best to
> only remove the "(2m)" in line 37.
> - I would also prefer if the comment were as neutral as possible, i.e.
> please remove the "rather strange" adjective in line 48.
> - not sure what a "positive check" is (compared to what?), but just
> saying "Therefore we start the VM with 4M heap." would be fine.
Oh, it was my fault. I mixed this test up with another one, which
comprises both positive and negative checks.
Fixed as you suggested.
Thanks a lot for verifying me.
More information about the hotspot-gc-dev