Young GC pause time definitions

Y. Srinivas Ramakrishna y.s.ramakrishna at oracle.com
Wed Jul 13 04:56:37 UTC 2011


On 7/12/2011 9:45 PM, Krystal Mok wrote:
> On Wed, Jul 13, 2011 at 1:28 AM, Y. Srinivas Ramakrishna<
> y.s.ramakrishna at oracle.com>  wrote:
>>
>>>
>>> The biggest change in our software is that we have reimplemented a big
>>> part of our main execution to rely on blocking queues (from java util
>>> concurrent) instead of an in-house queue implementation based on java
>>> monitors. I've had some wild guesses that synchronization on blocking
>>
>> Note problem with LinkedBlockingQueues (i'll dig up the CR id) that caused
>> GC behaviour similar to what you describe. Fixed in a later JDK (i'll dig
>> up
>> the details for you).
>
>
> Is it this one: http://bugs.sun.com/view_bug.do?bug_id=6805775
> Bug ID 6805775: LinkedBlockingQueue Nodes should unlink themselves before
> becoming garbage

Yes, thanks Kris, that's the one I had in mind. Given that it's indicated
as fixed in 6u19 and Christian is on 6u14, it's worth trying a JDK >= 6u19
and preferably the latest (6u26 or 6u27?).

-- ramki
_______________________________________________
hotspot-gc-use mailing list
hotspot-gc-use at openjdk.java.net
http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use



More information about the hotspot-gc-dev mailing list