Java VM segfaulting and other such strangeness

[plaz]

Hey folks,

I've been working on setting up Open-Xchange (http://www.open-xchange.org/), However I've noticed some very strange behaviour from the JVM.

After starting Tomcat and the three Open-Xchange components, it takes between 5 and 10 minutes before they bind to their ports and start accepting connections. During that time, there are a whole lot of Java processes running, but none are using any CPU time.

Once it's started, if I log into the Open-Xchange Groupware, the Java VM handling the Groupware server will jump to 100% CPU for several minutes, and then crash. It produced this dump:

# An unexpected error has been detected by HotSpot Virtual Machine:
#
# SIGSEGV (0xb) at pc=0x00000003, pid=724, tid=16386
#
# Java VM: Java HotSpot(TM) Server VM (1.5.0-b64 mixed mode)
# Problematic frame:
# C 0x00000003
Continued here (http://plaz.net.nz/open-xchange/jvmcrash.txt).

That's not the Java code crashing; it's an internal segfault in the Java virtual machine.

Has anyone noticed anything similar when running Java apps in the VPSes?

Cheers,

Sam

[retep]

Hi Sam.

I've found that virtually all crashes are due to one of two things.

1) not enough memory.

2) seen in some cases when running on a 2.6 kernel (hopefully this will be something that is resolved in the next few UML kernel releases).

Would you put in a trouble ticket for this one so we can log in and take a look?