Why does JVM 1.6.0_26 crash shortly after startup with SIGSEGV?

Frank Thiergart

Hey,

for the last year and a half I have maintained a small (20 User) confluence installation within an entreprise.
Just recently the VirtualMachine the server which the confluence instance was running on experienced a kernel panic.
I cannot give any info about the nature of this panic since I'm only the admin of my small virtual confluence server.

Shortly (I do not know if it was right after the complete shutdown) afterwards I got informed that the wiki wasn't running anymore.
I investigated the issue and found out that the JVM had crashed due to a SIGSEGV error.
So the server was restarted and confluence ran for 5-10 minutes.
However, the error reoccured. Neither an increase of available memory (from 1 to 2 gb) nor a completely new installation of all components solved this issue.

So these are the server specs: Server: VMWare Debian6 2.6.32-6-pve 32Bit 2gb ram.
Confluence: 3.5.1 with sun-java6-jdk 1.6.0_26 (1.7 was also tested)
DB: PostgreSQL

Here are the errors as in the catalina.out log:
Although it is stated here I used openjdk the same error applies when using java-jdk!!
#
# A fatal error has been detected by the Java Runtime Environment:
#
# SIGSEGV (0xb) at pc=0x012ef67d, pid=12927, tid=3063028592
#
# JRE version: 6.0_18-b18
# Java VM: OpenJDK Client VM (14.0-b16 mixed mode, sharing linux-x86 )
# Derivative: IcedTea6 1.8.9
# Distribution: Debian GNU/Linux 6.0 (squeeze), package 6b18-1.8.9-0.1~squeeze1
# Problematic frame:
# VException in thread "com.mchange.v2.async.ThreadPoolAsynchronousRunner$PoolThread-#0" [libjvm.so+0x34167d]
#
# An error report file with more information is saved as:
# /usr/local/confluence/confluence-3.5.1-std/hs_err_pid12927.log
java.lang.IllegalMonitorStateException
at java.lang.Object.wait(Native Method)
at com.mchange.v2.async.ThreadPoolAsynchronousRunner$PoolThread.run(ThreadPoolAsynchronousRunner.java:534)
#
# If you would like to submit a bug report, please include
# instructions how to reproduce the bug and visit:
# http://icedtea.classpath.org/bugzilla<br< a=""> />#

AND

#
# A fatal error has been detected by the Java Runtime Environment:
#
# Internal Error (nmethod.cpp:2175), pid=15452, tid=3016321904
# guarantee(nm->_lock_count >= 0) failed: unmatched nmethod lock/unlock
#
# JRE version: 6.0_26-b03
# Java VM: Java HotSpot(TM) Client VM (20.1-b02 mixed mode, sharing linux-x86 )
# An error report file with more information is saved as:
# /bin/hs_err_pid15452.log
#
# If you would like to submit a bug report, please visit:
# http://java.sun.com/webapps/bugreport/crash.jsp<br< a=""> />#


I also played around with the setenv.sh JAVA_OPTS values (I thought it was a memory problem first)
Here are my current values:
JAVA_HOME="/usr/lib/jvm/java-6-sun"
export JAVA_HOME
JAVA_OPTS="-Xms1024m -Xmx1024m -XX:MaxPermSize=256m $JAVA_OPTS -Djava.awt.headless=true "
export JAVA_OPTS

So if any of you have any idea please let me know.
I followed some suggestions in the help section, but those problems where like years away and are fixed in current versions of java and confluence.

Thanks in advance

PS: No other service is running on the JVM

Radu Dumitriu

OpenJDK is not very stable. You should use Sun^H^H^HOracle JDK.

SEGMENTATION FAULT is usually the consequence of accessing memory with an invalid address (there may be other causes). Usually there is an uninitialized pointer, a programming mistake anyway.

Open JDK is not supported by Atlassian.