SIGSEGV Segmentation Fault JVM Crash

Still need help?

The Atlassian Community is here for you.

Ask the community

Symptoms

Confluence (and other applications running on the JVM) become unavailable. Catalina.out or logs in the /var/log/messages may contain:

# # An unexpected error has been detected by HotSpot Virtual Machine: # # SIGSEGV (0xb) at pc=0xfe9bb960, pid=20929, tid=17 # # Java VM: Java HotSpot(TM) Server VM (1.5.0_01-b08 mixed mode) # Problematic frame: # V [libjvm.so+0x1bb960\] #   T H R E A D   Current thread (0x01a770e0): JavaThread "JiraQuartzScheduler_Worker-1" [_thread_in_vm, id=17\]  siginfo:si_signo=11, si_errno=0, si_code=1, si_addr=0x00000000  Registers: O0=0xf5999882 O1=0xf5999882 O2=0x00000000 O3=0x00000000 O4=0x00000000 O5=0x00000001 O6=0xc24ff0b0 O7=0x00008000 G1=0xfe9bb80c G2=0xf5999a48 G3=0x0a67677d G4=0xf5999882 G5=0xc24ff380 G6=0x00000000 G7=0xfdbc3800 Y=0x00000000 PC=0xfe9bb960 nPC=0xfe9bb964 ....

The following may appear in the atlassian-confluence.log:

[ERROR\] Thu Apr 08 11:43:05 PDT 2010 [com.atlassian.renderer.v2.V2Renderer\] Unable to render content due to system error: null  Throwable: java.lang.StackOverflowError at java.lang.ClassLoader.defineClass1(Native Method) at java.lang.ClassLoader.defineClass(ClassLoader.java:698) at java.security.SecureClassLoader.defineClass(SecureClassLoader.java:124) at org.apache.catalina.loader.WebappClassLoader.findClassInternal(WebappClassLoader.java:1817)

Diagnosis

  1. Check to confirm that Java has itself crashed.

    Linux
    ps -aux | grep java
    
    Windows
    Hit ctrl-alt-delete and check for java.exe in the running programs.
    
  2. Search the system for files named hs_err_pid*. Otherwise, look in /var/log/messages for a segmentation fault message.
    The SIGSEGV message indicates Java itself is crashing.

Cause

This is usually caused by a bug in the JVM, but in some cases, a Stack Overflow can cause the outage.

Resolution

  1. This is most likely due to the excerpt-include macro running recursively. See CONF-15247 for a resolution. This bug is fixed in Confluence 3.2. Upgrade to 3.2 or later.
  2. A secondary likely cause is a bug in the C2 (Server) compiler shipped in the Oracle 1.6 JDK for versions 18 and 19 (at the time of writing). This bug causes the JVM to exhaust the address space of the process. This affects both 32bit and 64bit versions of the JVM. Until a fix is delivered by Oracle, versions of the 1.6 JDK later than 17 should not be used to deploy Confluence. See Confluence performance / installation issue when using Java 6 Update 18 64 bit. Use JDK version 1.6.0_16.

If you have disabled the excerpt-include and excerpt macros as described in the workaround in CONF-15247 and have changed your JDK version, here are the next steps:

  1. Download the latest JDK and use it. Crashes are often caused by JDK bugs that get fixed over time.
  2. Ensure that your operating system is one of the Supported Platforms, and (especially for Solaris) ensure that the operating system has all current required patches applied.
  3. If running on a Virtual Machine (VM) move to a physical server and check to see if you can reproduce the problem (the bug may be specific to your VM).
  4. If you are using a 64bit JVM, try switching to 32bit. When doing so reduce your memory to -Xmx2048m as you cannot allocate more than this to a 32bit JVM.
  5. Follow the suggestions in Garbage Collector Performance Issues.
  6. Check CONF-19049 for further information and user contributions.

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

Last modified on Mar 30, 2016

Was this helpful?

Yes
No
Provide feedback about this article
Powered by Confluence and Scroll Viewport.