0

At work, I am trying to install Matlab Distributed Computing Server R2011a on a Rocks CentOS Cluster. Following the instructions, I mount the ISO and run ./install &. The installer shows a splash screen but then crashes and outputs the error message below.

A fatal error has been detected by the Java Runtime Environment:

SIGSEGV (0xb) at pc=0x0000003d39471c7c, pid=29673, tid=1105340736

JRE version: 6.0_17-b04 Java VM: Java HotSpot(TM) 64-Bit Server VM (14.3-b01 mixed mode linux-amd64 ) Problematic frame: C [libc.so.6+0x71c7c]

If you would like to submit a bug report, please visit: http://java.sun.com/webapps/bugreport/crash.jsp

My boss and I found an article at http://greg.porter.name/wiki/HowTo:MatlabOnRocks by someone who has the exact same problem, but never figured out how to fix it. Has anyone else encountered this problem? If anyone has found a fix, I would be very interested in knowing what steps you took to fix this problem.

Zhehao Mao
  • 1,789
  • 13
  • 13

2 Answers2

0

Sounds like a "try to run things on an unsupported platform"-issue.

In this kind of problems, ensure that you do this on a platform explicitly stated to support this combination. I do not believe that CentOS is a supported Oracle JVM platform.

Thorbjørn Ravn Andersen
  • 73,784
  • 33
  • 194
  • 347
  • That is entirely possible, but unfortunately we cannot avoid using CentOS. We are using the rocks cluster distribution, which uses a modified version of CentOS (and apparently a rather old version, since the default version of python is 2.4!). – Zhehao Mao Jun 03 '11 at 14:35
  • In that case ask MathLab about a version compiled for your system. – Thorbjørn Ravn Andersen Jun 03 '11 at 15:38
  • We ended up reverting back to R2010b and it worked. I'm going to mark this as the correct answer. – Zhehao Mao Jun 18 '11 at 13:17
0

It seem to be a general problem of java on memory limits set by cluster manager. We are encounter this problem with SGE on Fedora Core 13. see bug on fedora Solution is to set limits to ~ 4 Go (not easy all the time ...) or set environment

Guillermo
  • 1