Wednesday, 2 April 2014

Ooops, will I ever learn ? Problems with IBM Business Monitor 8.0.1.2 due to Linux libfreebl3.so library

Saw this whilst starting the Support cluster as part of an IBM Business Monitor 8.0.1.2 installation ( a fresh install using Oracle 11g R2 instead of my normal DB2 LUW ): -

native_stderr.log

JVMJ9VM135W /proc/sys/kernel/core_pattern setting "|/usr/libexec/abrt-hook-ccpp %s %c %p %u %g %t e" specifies that core dumps are to be piped to an external program. The JVM may be unable to locate core dumps and rename them.
java.lang.UnsatisfiedLinkError: Ws60ProcessManagement (/opt/IBM/WebSphere/AppServer/profiles/AppSrv01/cognos/BAM8012.Support.AppSrv01Node.0/bin64/libfreebl3.so: version `NSSRAWHASH_3.12.3' not found (required by /lib64/libcrypt.so.1))
        at java.lang.ClassLoader.loadLibraryWithPath(ClassLoader.java:1091)
        at java.lang.ClassLoader.loadLibraryWithClassLoader(ClassLoader.java:1055)
        at java.lang.System.loadLibrary(System.java:510)
        at com.ibm.ws.natv.util.LibraryLoader.loadLibrary(LibraryLoader.java:52)
        at com.ibm.ws.process.ProcessLibraryLoader.loadDistributedLibrary(ProcessLibraryLoader.java:45)
        at com.ibm.ws.process.ProcessLibraryLoader.loadNativeLibrary(ProcessLibraryLoader.java:36)
        at com.ibm.ws.process.linuxutil.ThreadUtil.<clinit>(ThreadUtil.java:46)
        at java.lang.J9VMInternals.initializeImpl(Native Method)
        at java.lang.J9VMInternals.initialize(J9VMInternals.java:237)
        at com.ibm.ws.process.linuxutil.Mutex.<init>(Mutex.java:60)
        at com.ibm.ws.process.LinuxProcessImpl.<init>(LinuxProcessImpl.java:84)
        at java.lang.J9VMInternals.newInstanceImpl(Native Method)
        at java.lang.Class.newInstance(Class.java:1563)


The solution ?

Read my own notes. Read my own blog posts: -


and remember to set the required Linux LD_PRELOAD variable: -

echo export LD_PRELOAD=/lib64/libfreebl3.so >> /home/wasadmin/.bashrc

This modifies the environment for the wasadmin user, and required me to terminate the various WAS JVMs ( all of which had been initiated by that user ) and restart my Linux session ( log out and back in again ).

Once I did this, and restarted the Deployment Manager and Node Agent, the Support cluster started up like a dream :-)

No comments: