I am finding my Solaris 10 /tmp directory is rapidly filling up with error logs in the form of hs_err_pidXXXX.log and is being generated every few minutes. I have heard somewhere (I do not recall) that disableing the service `svcadm disable svc:/system/webconsole:console` and rebooting the system will help - but these darn error logs just keep piling up. I've seen listings with differing topics pointing to error code that to me all looks the same, but nobody has a suggested workaround or a fix.
I am lost - can anybody help?
Error in total is as follows:
# An unexpected error has been detected by HotSpot Virtual Machine:
# SIGSEGV (0xb) at pc=0xfe9b4124, pid=233, tid=1
# Java VM: Java HotSpot(TM) Server VM (1.5.0_18-b02 mixed mode)
# Problematic frame:
# V [libjvm.so+0x1b4124]
--------------- T H R E A D ---------------
Current thread (0x000387d8): JavaThread "main" [_thread_in_vm, id=1]
Well, your system is experiencing segmentation violations (i.e. memory accessing errors) probably due to some errant pointer in some native code, or due to some soft/corrupt memory in your system.
The errant pointer problem is the most likely but I have seen cases where memory was getting corrupted in some embedded systems and caused similar sort of problems. However, those faults appeared in the pthreads library and not in the VM itself.
Your memory access problems are in libjvm which is very worrying. This is usually solid as a rock. If it wasn't Java would be unreliable.
If a later version of the JVM is available I would upgrade to that and see if the problem persists.
The fact that your system is rapidly filling up with these error logs would make me think about running some kind of extensive memory check on your system. When I have seen soft/corrupted memory on my embedded systems it has been intermittent. The fact that you see the problem so often would tend to make it more likely that you have some kind of memory corruption going on. So, before going any further, I would run that system memory test. If the memory comes up clean it mean that something [awful] is happening in your JVM.