RFR: 8012241 Kitchensink crashed with OOME on linux x64
zhengyu.gu at oracle.com
Wed Jun 19 05:27:20 PDT 2013
Personally, I have yet to reproduce, but we see a few from Aurora
On 6/18/2013 2:20 PM, Andrew Haley wrote:
> On 06/18/2013 07:10 PM, Zhengyu Gu wrote:
>> NMT shuts itself down when it encounters OOM. But most of time, OOM is encountered by other parts of VM, which could result in fatal error. The change is to intercept OOM in os::malloc() and shutdown NMT, hopefully will free enough memory to allow VM to continue executing.
>> Bug: https://jbs.oracle.com/bugs/browse/JDK-8012241
>> External bug: Not available
>> Webrev: http://cr.openjdk.java.net/~zgu/8012241/webrev.00/
> Was it actually possible to test this?
More information about the hotspot-runtime-dev