JVM crashes during deoptimization phase

Linux version:
 Linux version 2.4.9-e.3smp (###@###.###) (gcc
 version 2.96 20000731 (Red Hat Linux 7.2 2.96-108.1)) #1 SMP Fri May 3
 16:48:54 EDT 2002 - RedHat AS 2.1 - hyperthreading was on.



After a 14 hour run, the memory on the machine was about
5M free (machine has 2Gb and the jvm heap is set to 1024M) - the JVM
crashed a few minutes ago (traffic was still running) with the following
error. The verbose gc log in attachment which also has the the crash report:

-----------------------------------------------------------------------
53479.741: [GC 53479.741: [ParNew: 16255K->0K(16320K), 0.0771970 secs]
507744K->497333K(1048512K), 0.0773430 secs]
53480.563: [GC 53480.563: [ParNew: 16255K->0K(16320K), 0.0914510 secs]
513589K->502901K(1048512K), 0.0915960 secs]
53481.733: [GC 53481.733: [ParNew: 16255K->0K(16320K), 0.0889060 secs]
519157K->508893K(1048512K), 0.0890670 secs]
#
# HotSpot Virtual Machine Error, Internal Error
# Please report this error at
# http://java.sun.com/cgi-bin/bugreport.cgi
#
# Java VM: Java HotSpot(TM) Server VM (1.4.2_02-b03 mixed mode)
#
# Error ID: 434F444523414348450E43505000CF
#
# Problematic Thread: prio=1 tid=0x8fce1220 nid=0x2102 runnable
#

Heap at VM Abort:
Heap
  par new generation total 16320K, used 6086K [0x44b20000, 0x45b20000,
0x45b20000)
   eden space 16256K, 37% used [0x44b20000, 0x4512b0a8, 0x45b00000)
   from space 64K, 0% used [0x45b00000, 0x45b00000, 0x45b10000)
   to space 64K, 0% used [0x45b10000, 0x45b10000, 0x45b20000)
  concurrent mark-sweep generation total 1032192K, used 508893K
[0x45b20000, 0x84b20000, 0x84b20000)
  concurrent-mark-sweep perm gen total 34188K, used 24914K [0x84b20000,
0x86c83000, 0x88b20000)


The "error id" 434F444523414348450E43505000CF
decodes to: codeCache.cpp, 207


--------------------------------------------------