Puzzling - why is a Full GC triggered here?

[Krystal Mok]

Hi,

I had a BTrace script that could trace System.gc() calls and print the
stack trace [1]. If you run the script with a BTrace agent from the start
of your program, you might be able to track down the caller [2]

HTH,
- Kris

[1]:
https://gist.github.com/2000950/37b0095b1edbb7d4a43cc5b39cbe148c8184d3aa#file_trace_system_gc_call.java
[2]: http://kenai.com/projects/btrace/pages/UserGuide

On Sat, Mar 31, 2012 at 4:32 AM, Kobe Bryant <kbbryant61 at gmail.com> wrote:

> The following GC trace was obtained soon after the JVM was started
> (a lot of transient objects are generated as a part of the start up and
> hence
> a lot of minor GCs).
>
> ( Xms = Xmx = 2G, NewGen = 1/3 of 2G, PermGen = 760m)
>
> I see Full GC triggered (highlighted in bold below): the object space
> occupancy
> in tenured space is 2% - and yet a Full GC is triggered (I have elided
> PermGen reports, as PermGen usage is about 2% and is healthy)
>
> {Heap before GC invocations=2 (full 1):
>  PSYoungGen      total 608256K, used 2016K [0x00000007d6c00000,
> 0x0000000800000000, 0x0000000800000000)
>   eden space 540672K, 0% used
> [0x00000007d6c00000,0x00000007d6c00000,0x00000007f7c00000)
>   from space 67584K, 2% used
> [0x00000007f7c00000,0x00000007f7df8030,0x00000007fbe00000)
>   to   space 67584K, 0% used
> [0x00000007fbe00000,0x00000007fbe00000,0x0000000800000000)
>  PSOldGen        total 1421312K, used 0K [0x0000000780000000,
> 0x00000007d6c00000, 0x00000007d6c00000)
>   object space 1421312K, 0% used
> [0x0000000780000000,0x0000000780000000,0x00000007d6c00000)
>  PSPermGen       total 524288K, used 10939K [0x0000000750400000,
> 0x0000000770400000, 0x0000000780000000)
>   object space 524288K, 2% used
> [0x0000000750400000,0x0000000750eaef28,0x0000000770400000)
> *2012-03-26T22:35:49.121-0700: 0.614: [Full GC (System)AdaptiveSizeStart:
> 0.645 collection: 2*
> AdaptiveSizeStop: collection: 2
>  [PSYoungGen: 2016K->0K(608256K)] [PSOldGen: 0K->1809K(1421312K)]
> 2016K->1809K(2029568K) [PSPermGen: 10939K->10939K(524288K)], 0.0317230
> secs] [Times: user=0.03 sys=0.00, real=0.03 secs]
>
> Why was full GC triggered here? Is it because RMI DGC? Is there a way to
> identify if DGC is triggering GCs?
> PermGen is at a healthy and consistent level of 2%. I have hence elided
> this detail above. And hence,
> this cannot be the cause of full GCs.
>
> thanking you
>
>
>
> _______________________________________________
> hotspot-gc-use mailing list
> hotspot-gc-use at openjdk.java.net
> http://mail.openjdk.java.net/mailman/listinfo/hotspot-gc-use
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.openjdk.java.net/pipermail/hotspot-gc-use/attachments/20120331/7e73b2a4/attachment.html