Determine the method optimizing which the Jrockit JVM crashed.

Question

I have an application that is running on Jrockit JVM. which crashed after 3-4 hours.
From the jrockit.<pid>.dump I could determine that the crash was on the Thread: "(Code Optimization Thread 1) .
And by using -XnoOpt we can stop code Optimization.
However, how do we determine which exact method caused it so that i can use the option -XX:OptFile
I am running Jrockit R28.0.1-21-133393-1.6.0_20-20100512-2126-linux-x86_64
on RED HAT LINUX 5.3 (64 bit)
Thread:
"(Code Optimization Thread 1)" id=5 idx=0x54 tid=29997 lastJavaFrame=0xfffffffffffffffc
Stack 0: start=0x41a70000, end=0x41ab2000, guards=0x41a75000 (ok), forbidden=0x41a73000
Thread Stack Trace:
at mspace_free+473(osal_mspace.c:4608)
at irAliasValidate+1760(aliases.c:558)
at irInfoGet+61(irinfo.c:143)
at irAliasMustBeAliases+59(aliases.c:112)
at update_callvector+4976(inline.c:284)

Answer 1

Many bugs were reported for JVM crast at acGetOperand in JRockit R27.x
But no bugs are reported exactly with the same stack trace.
Regarding a workaround, the below stack is actually misleading us. It does not give us a clue about the reasons for the crash (whether it is happening due to concurrent GC or parallel GC or Compaction etc.), so we cannot really say what GC settings would avoid it.
I recommend that we collect another textual dump if the crash occurs again and hope that the stack trace generated the next time will give us better insight into the root cause.
Another suggestion would be to upgrade to R28.x