I've got the below error from WebSphere Application Server's SystemOut.log file:
[5/11/12 8:13:46:291 GMT+08:00] 000063c0 WebApp E [Servlet Error]-[ICMResourceManager]: java.lang.OutOfMemoryError: unable to allocate 79683 bytes for native buffer
at java.io.FileOutputStream.writeBytes(Native Method)
at java.io.FileOutputStream.write(FileOutputStream.java:290)
at com.ibm.mm.icmrm.storage.RMOutputStream.write(RMOutputStream.java:65)
at com.ibm.mm.icmrm.storage.RMDataTransfer.copyData(RMDataTransfer.java:150)
at com.ibm.mm.icmrm.storage.filesystem.FileSystem.store(FileSystem.java:104)
It seems the native OOM issue. I have a question here, is it the 'native buffer' here same with 'direct byte buffer'? Can I try to enlarge 'native buffer' size by reset property XX:MaxDirectMemorySize? BTW, what's the default value for it?
Or can I set this property -Dcom.ibm.nio.DirectByteBuffer.AggressiveMemoryManagement=true to make native memory space dynamic?
WAS platform information:
WebSphere Platform 6.1 [ND 6.1.0.27 cf270937.16] running with process name cmrmCell01\cmrmNode01\icmrm1 and process id 1802306
Detailed IFix information: Please use the versionInfo command to view this information
Host Operating System is AIX, version 5.3
Java version = 1.5.0, Java Compiler = j9jit23, Java VM name = IBM J9 VM
I know this isn't much of an answer, but see if this article, Understanding how the JVM uses native memory on AIX, has anything useful.
Related
Iam facing JVM Crash cosistently while enabling hotdeploy (USING below java options on starting up JAVA_OPTS -Xmx4096m -XX:MetaspaceSize=512m -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=crash -XX:ThreadStackSize=512 -XX:+UseConcMarkSweepGC -XX:ParallelGCThreads=5 -XX:NewRatio=2 -XX:+UnlockDiagnosticVMOptions -XX:-UseLoopPredicate -Xdebug -Xrunjdwp:transport=dt_socket,address=$DEBUG_PORT,server=y,suspend=n -XX:NewRatio=2 -Dspringloaded.synchronize=true JAVA_OPTS=`echo $JAVA_OPTS -Dspringloaded.synchronize=true -javaagent:springloaded-1.2.1.jar -noverify
)
Environment : JDK 1.8 U 66, RHEL 6.7
#
# A fatal error has been detected by the Java Runtime Environment:
#
# SIGSEGV (0xb) at pc=0x00007faee9a1e27c, pid=27208, tid=140379827795712
#
# JRE version: Java(TM) SE Runtime Environment (8.0_66-b17) (build 1.8.0_66-b17)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.66-b17 mixed mode linux-amd64 )
# Problematic frame:
# V [libjvm.so+0x35027c] Canonicalizer::do_If(If*)+0x1c
#
# Core dump written. Default location: core.27208
#
# An error report file with more information is saved as:
# hs_err_pid27208.log
# [ timer expired, abort... ]
I've noticed both -javaagent and -noverify in Java options list.
It looks like springloaded agent generates invalid bytecode, while the bytecode verification is explicitly turned off. No surprise, this may lead to unpredictable results including JVM crash.
This is not a JVM problem, but most likely a bug in springloaded agent. Try to remove -noverify option.
-XX:-TieredCompilation may also work around this particular problem, but don't expect application to work correctly if the bytecode fails to pass verification. It's better to stay away from the buggy agent libraries.
4.2.1 Crash in HotSpot Compiler Thread or Compiled Code
If the fatal error log indicates that the crash occurred in a compiler
thread, then it is possible (but not always the case) that you have
encountered a compiler bug. Similarly, if the crash is in compiled
code then it is possible that the compiler has generated incorrect
code.
In the case of the HotSpot Client VM (-client option), the compiler
thread appears in the error log as CompilerThread0. With the HotSpot
Server VM there are multiple compiler threads and these appear in the
error log file as CompilerThread0, CompilerThread1, and AdapterThread.
Below is a fragment of an error log for a compiler bug that was
encountered and fixed during the development of J2SE 5.0. The log file
shows that the HotSpot Server VM is used and the crash occurred in
CompilerThread1. In addition, the log file shows that the Current
CompileTask was the compilation of the java.lang.Thread.setPriority
method.
An unexpected error has been detected by HotSpot Virtual Machine:
:
Java VM: Java HotSpot(TM) Server VM (1.5-internal-debug mixed mode) :
--------------- T H R E A D ---------------
Current thread (0x001e9350): JavaThread "CompilerThread1" daemon
[_thread_in_vm, id=20]
Stack: [0xb2500000,0xb2580000), sp=0xb257e500, free space=505k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code,
C=native code) V [libjvm.so+0xc3b13c] :
Current CompileTask: opto: 11 java.lang.Thread.setPriority(I)V
(53 bytes)
--------------- P R O C E S S ---------------
Java Threads: ( => current thread ) 0x00229930 JavaThread "Low
Memory Detector" daemon [_thread_blocked, id=21]
=>0x001e9350 JavaThread "CompilerThread1" daemon [_thread_in_vm, id=20] :
In this case there are two potential workarounds:
The brute force approach: change the configuration so that the application is run with the -client option to specify the HotSpot
Client VM.
Assume that the bug only occurs during the compilation of the setPriority method and exclude this method from compilation.
The first approach (to use the -client option) might be trivial to
configure in some environments. In others, it might be more difficult
if the configuration is complex or if the command line to configure
the VM is not readily accessible. In general, switching from the
HotSpot Server VM to the HotSpot Client VM also reduces the peak
performance of an application. Depending on the environment, this
might be acceptable until the actual issue is diagnosed and fixed.
The second approach (exclude the method from compilation) requires
creating the file .hotspot_compiler in the working directory of the
application. Below is an example of this file:
exclude java/lang/Thread setPriority
In general the format of this file is exclude CLASS METHOD, where
CLASS is the class (fully qualified with the package name) and METHOD
is the name of the method. Constructor methods are specified as
and static initializers are specified as .
Note - The .hotspot_compiler file is an unsupported interface. It is
documented here solely for the purposes of troubleshooting and finding
a temporary workaround.
Once the application is restarted, the compiler will not attempt to
compile any of the methods listed as excluded in the .hotspot_compiler
file. In some cases this can provide temporary relief until the root
cause of the crash is diagnosed and the bug is fixed.
In order to verify that the HotSpot VM correctly located and processed
the .hotspot_compiler file that is shown in the example above, look
for the following log information at runtime. Note that the file name
separator is a dot, not a slash.
Excluding compile: java.lang.Thread::setPriority
Source
Agree with #apangin, In the program you are doing bytecode intrumentation (-agent) but specifies -noverify. When verification is turned off, you may end up such crashes.
You should not use -noverify or -Xverify:none during byte code intrumentation.
For those of you unfamiliar with bytecode verification, it is simply part of the JVM's classloading process that checks the code for certain dangerous and disallowed behavior. You can (but shouldn't) disable this protection on many JVMs by adding -Xverify:none or -noverify to the Java command line. https://blogs.oracle.com/buck/entry/never_disable_bytecode_verification_in
My WebLogic Servers use JRockit JVM R28. We need to have the WebLogic JVMs configured to automatically shutdown/kill/exit when an OutOfMemoryError occurs.
A JRockit JVM Parameter called "ExitOnOutOfMemory" will let us accomplish this. However Oracle documentation provides incorrect and conflicting information.
1.) http://docs.oracle.com/cd/E13150_01/jrockit_jvm/jrockit/jrdocs/refman/optionXX.html says to simply put "-XXexitOnOutOfMemory" into startup scripts. However, JRockit doesnt "recognize" this parameter.
2.) http://docs.oracle.com/cd/E15289_01/doc.40/e15062/optionxx.htm#BABCDAIB says to put "-XX:+ExitOnOutOfMemoryError" into startup scripts. However JRockit does not recognize this configuration either. I believe they mistakenly copied this from Hotspot documentation.
How do i implement this parameter?
-XX:+ExitOnOutOfMemoryError works as expected with JRockit R28.2.2:
$ jrockit-jdk1.6.0_29/bin/java -Xmx20m -XX:+ExitOnOutOfMemoryError OOM
java.lang.OutOfMemoryError: allocLargeObjectOrArray: [B, size 40976
at jrockit/vm/Allocator.allocLargeObjectOrArray(JIZ)Ljava/lang/Object;(Native Method)
at jrockit/vm/Allocator.allocObjectOrArray(Allocator.java:349)
at jrockit/vm/Allocator.allocArray(Allocator.java:257)
at OOM.<init>(OOM.java:3)
at OOM.main(OOM.java:9)
at jrockit/vm/RNI.c2java(JJJJJ)V(Native Method)
-- end of trace
[ERROR] Exit on OutOfMemory requested. Exiting.
JRockit aborted: Exit requested on OOM (51)
Which version of JRockit are you using? Did you spell the parameter correctly?
we are using NIOFSDirectory instead of MMapDirectory in an intensive test load for indexing. But we are still getting a SIGSEV with MMapDirectory. Is there anything else I should be looking for? This is a Linux 64 bit machine. Head of the stack trace follows:
Please let me know what more info would be helpful.
Thanks...
A fatal error has been detected by the Java Runtime Environment:
SIGSEGV (0xb) at pc=0x00002aaaab285ea7, pid=2408, tid=1128642880
JRE version: 6.0_25-b06
Java VM: Java HotSpot(TM) 64-Bit Server VM (20.0-b11 mixed mode linux-amd64 compressed >oops)
Problematic frame:
J org.apache.lucene.store.MMapDirectory$MMapIndexInput.readByte()B
You are most likely closing your IndexReader while other threads are still using it?
Lucene tries to catch this mis-use but it's detection is best-effort only.
You could also try calling MMapDirectory.setUseUnmap(false) to turn off unmapping; this should prevent the SIGSEGV.
I want to increase the heap size of my Glassfish. For that, I know that I can go up to 4GB:
java -Xmx4000M -version
java version "1.6.0_26"
Java(TM) SE Runtime Environment (build 1.6.0_26-b03-384-10M3425)
Java HotSpot(TM) 64-Bit Server VM (build 20.1-b02-384, mixed mode)
I tried to set in domain.xml file the -Xmx to 2GB:
<jvm-options>-Xmx2000m</jvm-options>
But I am getting the following error:
asadmin> start-domain
Waiting for ... to start .Error starting domain ...
The server exited prematurely with exit code 1.
Before it died, it produced the following output:
Error occurred during initialization of VM
The size of the object heap + VM data exceeds the maximum representable size
launchctl bsexec failed: Inappropriate ioctl for device
Launching the command with -v option gives this:
12 oct. 2011 11:46:34 com.sun.enterprise.admin.launcher.GFLauncherLogger info
INFO: JVM invocation command line:
/System/Library/Java/JavaVirtualMachines/1.6.0.jdk/Contents/Home/bin/java
-XX:+UnlockDiagnosticVMOptions
-XX:MaxPermSize=512m
-XX:NewRatio=2
-XX:+CMSClassUnloadingEnabled
-Xmx2000m
-Xms1000m
...
12 oct. 2011 11:46:35 com.sun.enterprise.admin.launcher.GFLauncherLogger info
INFO: Successfully launched in 45 msec.
Error occurred during initialization of VM
Could not reserve enough space for object heap
Could not create the Java virtual machine.
Command start-domain failed.
I found the problem. For some reasons which I do not understand, the "-d32" argument was passed in the command, asking to start in 32 bits mode. When adding
<jvm-options>-d64</jvm-options>
to the domain.xml file, glassfish starts. Please note that this option is not present by default in the file.
you must be using a different Java install when doing the "-version" and from within GlassFish. I can reproduce the same error with the default "java" in Mac OS' path but not when using an absolute path :
% /System/Library/Java/JavaVirtualMachines/1.6.0.jdk/Contents/Home/bin/java -Xmx2300m -version
java version "1.6.0_26"
Java(TM) SE Runtime Environment (build 1.6.0_26-b03-383-11A511)
Java HotSpot(TM) 64-Bit Server VM (build 20.1-b02-383, mixed mode)
% java -Xmx2300m -version
Error occurred during initialization of VM
Could not reserve enough space for object heap
Could not create the Java virtual machine.
Could it be that you don't have enough disk space on your system? Reserving 4GB of RAM may require MAC OS to expand swap space, and you may not have the disk space to do this. Unfortunately, this has burned me before :-/
I observe in consecutive javacore dumps that the addresses of classloaders are changing in Websphere 6.1 (IBM JVM 1.5.0). The classes below the classloader have unchanged address. So isn't it the same classloader at a new address?
E.g. in first javacore
2CLTEXTCLLOAD Loader com/ibm/ws/classloader/CompoundClassLoader(0x00002AAABF5BB7F0)
3CLTEXTCLASS org/eclipse/emf/ecore/EObject(0x00002AAB24684B30)
and in a later javacore
2CLTEXTCLLOAD Loader com/ibm/ws/classloader/CompoundClassLoader(0x00002AAABF5AB6E0)
3CLTEXTCLASS org/eclipse/emf/ecore/EObject(0x00002AAB24684B30)
Same class previously under a classloader at 0x00002AAABF5BB7F0, and now the classloader is at 0x00002AAABF5AB6E0
I'm trying to figure out a class unloading problem, and this does not seem to make life easier when digging through heap dumps.
Perhaps the GC is compacting the heap. Consider -Xnocompactgc or -Xnocompactexplicitgc as a generic JVM arguments. See IBM JDK diagnoses documentation:
http://www.ibm.com/developerworks/java/jdk/diagnosis/
If the application was stopped and restarted wouldnt' new classloaders be created for your classes because there is an application classloader?
Are you concerned about this because of native heap consumption or something else?