The J2EE application is experiencing random crashes. The crashes happen even when the application runs without load

The application is a J2EE application deployed into a set of instances of Weblogic AppServer. Hardware: E15 running Solaris 8 with 64GB of memory.

The JVM options used to run an instance of WLS are:

-server -XX:+AggressiveHeap -XX:-UseAdaptiveSizePolicy
-Xms3502m -Xmx3502m
-XX:MaxPermSize=128M

-Dsun.rmi.dgc.server.gcInterval=3600000 -Dsun.rmi.dgc.client.gcInterval=3600000
-verbose:gc -XX:+PrintGCTimeStamps -XX:+PrintGCDetails -XX:+PrintHeapAtGC -Xloggc:/var/tmp/mgd2_log_file_name.log


The application (an instance of WLS) crashes randomly: the crash may occur even without any user load/app activity, a few minutes after deployment. At times, no GC output could be taken as there were no GC activity prior to the crash. The error thrown by the WLS instance immediately before the crash is:

Exception in thread "CompilerThread1" java.lang.OutOfMemoryError: requested 32756 bytes for ChunkPool::allocate. Out of swap space?

The application is not using much of its memory resources:
The jvmstat's VisualGC showed plenty of heap space left: <50% of Eden space, <2% of Old Gen. space, and <50% of Perm. Gen.
 

When tried to set equal permanent generation limits(-XX:PermSize=128m -XX:MaxPermSize=128m), another kind of error was thrown:

Exception java.lang.OutOfMemoryError: requested 2621440 bytes for GrET* in /export1/jdk142-update/ws/fcs/hotspot/src/share/vm/utilities/growableArray.cpp. Out of swap space?


The application does not crash when heap memory is set to lower values, e.g. 3400m (both min and max).

The application does not crash when AggressiveHeap option is not used: using ParallelGC seems to work fine.



###@###.### 2005-2-16 16:17:23 GMT