>  INFO [GC inspection] 2010-07-21 01:01:49,661 GCInspector.java (line 110) GC for ConcurrentMarkSweep: 11748 ms, 413673472 reclaimed leaving 9779542600 used; max is 10873667584
> ERROR [Thread-35] 2010-07-21 01:02:10,941 CassandraDaemon.java (line 78) Fatal exception in thread Thread[Thread-35,5,main]
> java.lang.OutOfMemoryError: Java heap space
>  INFO [GC inspection] 2010-07-21 01:02:10,958 GCInspector.java (line 110) GC for ConcurrentMarkSweep: 10043 ms, 259576 reclaimed leaving 10172790816 used; max is 10873667584

So that confirms a "legitimate" out-of-memory condition in the sense
that CMS is reclaiming extremely little and the live set after a
concurrent mark/sweep is indeed around the 10 gig.


-- 
/ Peter Schuller

Reply via email to