[rules-users] PermGen space error

Joe Ammann joe at pyx.ch
Thu Feb 9 07:27:00 EST 2012


On 02/09/2012 09:34 AM, Geoffrey De Smet wrote:
> The perm gen space is a subset of the heap space.
> So it doesn't make sense to put them on the same value.
>
> Try this instead:
>     -Xmx1024m -XX:MaxPermSize=512m
> Or, if that fails:
>     -Xmx2048m -XX:MaxPermSize=1024m
>
Sorry to disagree Geoffrey, but PermGen is NOT a part of heap in all
(Sun) JVM's that I'm aware of. See

http://stackoverflow.com/questions/2129044/java-heap-terminology-young-old-and-permanent-generations
http://www.oracle.com/technetwork/java/gc-tuning-5-138395.html (the
picture a few pages down)

So the sizes are specified separately. But having a MaxPerm of 1G and
STILL see PermGen OOM is strange.

The problem could be, that when using the CMS (concurrent mark sweep)
GC, PermGen is not collected. I have seen cases where sooner or later
OOM PermGen errors happened consistently, until we added

|-XX:+CMSPermGenSweepingEnabled|

-- 
CU, Joe

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/rules-users/attachments/20120209/6c6188f5/attachment.html 


More information about the rules-users mailing list