A java.lang.OutOfMemoryErrors in the tracefile could mean
that there is a memory leak in the AdminServer Java Virtual Machine (JVM™)
process. However, this error might indicate the configured heap for the
AdminServer JVM is not large enough.
To increase the heap size of the AdminServer process, do the
following.
- Create backup copy of WebSphere/AppServer/bin/admin.config.
- Edit the admin.config file.
- Find the line
com.ibm.ejs.sm.util.process.Nanny.adminServerJvmArgs and change the
heap size from -Xmx128m (128 MB) to -Xmx256m.
- If increasing the -Xmx parameter does not prevent the
java.lang.OutOfMemoryError, enable -verbosegc for the AdminServer by
adding -verbosegc to the
com.ibm.ejs.sm.util.process.Nanny.adminServerJvmArgs.
- Stop and restart WebSphere AdminServer and attempt to recreate the
problem.
- The verbosegc information is written to the
WebSphere/AppServer/logs/tracefile file.
- Examine this information to determine if there is a memory leak in the
AdminServer. This is the case if the amount of heap memory in use, as
shown by the verboseGC output,is continually increasing.
- If there is a potential memory leak, use either the runhprof facility
or the IBM® heapdump to examine the contents of the heap and determine
what is leaking. For more information on how to enable the IBM heapdump,
refer to the following technotes.
- Refer to MustGather:
Readme First to locate instructions for capturing additional
information.
|